University of Wollongong Research Online
University of Wollongong Thesis Collection University of Wollongong Thesis Collections
2009 Mercury investigations in remote Oceania: wet deposition and bio-accumulation on Tutuila Island Peter Peshut University of Wollongong, [email protected]
Recommended Citation Peshut, Peter, Mercury investigations in remote Oceania: wet deposition and bio-accumulation on Tutuila Island, Doctor of Philosophy thesis, School of Earth and Environmental Sciences, University of Wollongong, 2009. http://ro.uow.edu.au/theses/3637
Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: [email protected]
MERCURY INVESTIGATIONS IN REMOTE OCEANIA: WET DEPOSITION AND BIO-ACCUMULATION ON TUTUILA ISLAND
A thesis submitted in partial fulfillment of the requirements for the award of the degree of
Doctor of Philosophy in Environmental Science
from
University of Wollongong
by
Peter Joseph Peshut, BSc., MSc.
School of Earth and Environmental Sciences Faculty of Science
October 2009 DECLARATION
I, Peter Joseph Peshut, hereby declare that this thesis, submitted in partial fulfillment of the requirements for the award of Doctor of Philosophy from the School of Earth and Environmental Sciences, University of Wollongong, is my own original composition and has not been submitted for qualifications at any other academic institution, and that all work was conducted by me or with assistance that was under my direct supervision, and that results presented are from my own original investigations unless otherwise acknowledged.
Peter Joseph Peshut 15 October 2009
ii ACKNOWLEDGEMENTS
The scope of this work ensures that no claim can be made for an independent accomplishment by the author. Recognition is gratefully given to the talented and dedicated individuals without whom this work would not have been attempted. In the field, Howard Dunham and crew of Moon Divers (Tutuila Island) un-failingly lived their motto “we’ll do the job” throughout two years of field initiatives, that involved driving small boats, diving, and technical assistance for collections, by day and by night, on dozens of forays of up to 90 km, in all sea conditions and weathers. In the laboratory, Brenda Lasorsa of Battelle Marine Sciences Laboratory (Sequim, Washington, USA) provided extensive guidance on field activities, ensured a steady flow of supplies, and maintained superior QA/QC for Hg analyses in 400 field samples from rainfall, sediments, surface waters, and biota. Edna Buchan, Water Program Manager, and her staff of the American Samoa Environmental Protection Agency (Tutuila Island) provided technical assistance and logistical support, and endless good cheer, frequently outside regular duty hours. For thesis production, all geographical figures were prepared from the American Samoa GIS database by Raymond Laine from the School of Earth and Environmental Sciences, University of Wollongong. Carl Goldstein of the United States Environmental Protection Agency (San Francisco, USA) provided administrative support, funding, and sustaining enthusiasm.
iii ABSTRACT
Mercury was investigated in multiple environmental compartments of the remote oceanic island of Tutuila in southern Polynesia, to examine prevalent lines of thought on how Hg is distributed in the environment at the global scale. Research shows that
Hg is highly mobile in the environment, and that the atmosphere is the principal reservoir from which Hg is ultimately deposited to aquatic systems. Since the poisonings at Minamata and Niigata fifty years ago, Hg is a highly studied element, and the neurotoxicity and accumulation of Hg in the biosphere at levels that pose risks to human health are well documented. The ubiquity of Hg in the global environment is not as well known. Direct experimental evidence for Hg in remote regions of the globe is rare, which hinders our understanding of the extent of Hg proliferation and weakens arguments that Hg is a contaminant of global concern. To assess the prevalence of Hg in the remote global environment, this study provided experimental evidence for Hg in rainfall and in coral reef biota, at a remote oceanic location thousands of kilometres from any Hg emission source. Other environmental compartments of soils, erosion materials, and reef sediments were investigated to examine Hg transport and distribution on this remote island where there was little previous data on the occurrence of Hg. Trace-elements field and analytical protocols were maintained throughout to ensure integrity of the data set. Speciation analysis was applied to most field samples to determine proportions of inorganic and organic Hg in environmental media, to examine potential sites of Hg methylation, and to accurately assess the prevalence of the highly toxic methylmercury compared to the relatively less toxic inorganic species. Mercury in Tutuila rainfall was the same, or slightly higher than, Hg in rainfall from comparably remote sites in the northern hemisphere, indicating that Hg is well-mixed and
iv distributed globally via the atmosphere. Mercury residues in top-trophic fish from
Tutuila reefs were far above allowable world health standards. Results for Hg in rainfall and predator fish support concerns for global proliferation of Hg, and are consistent with evidence that world-wide, Hg deposition has increased since the onset of industrialisation. Evaluation of transport and distribution of Hg on Tutuila strongly suggested that atmospheric wet deposition is a significant source of Hg to aquatic environments of this remote oceanic high island. Patterns of Hg transport and distribution indicated that the geologic base or regional volcanism were not likely to be significant sources of Hg on Tutuila. The potential for human habitation and a modern material economy as a non-point source of Hg to reef systems in remote global locations, due to Hg in consumer goods, vehicles, and building materials, was identified, though this appeared to be highly localised. Although Hg in rainfall and in top-trophic reef fish indicated that the flux of Hg to remote Tutuila is significant, Hg did not occur at levels above established regulatory or health standards in the other compartments investigated. Mercury in lower-trophic fish, reef sediments, and water column, at pristine sites do not yet appear to be significantly impacted by the globalisation of Hg. At the study site under the influence of anthropogenic disturbance, slightly elevated Hg was found in the water column, sediments, and fish, compared with pristine sites, but concentrations were still generally lower than established threshold levels of concern, except for a few benthic-feeding fish. Overall, study results supported that the proliferation of Hg via atmospheric processes is of global relevance.
Notwithstanding the weight of mounting evidence, international consensus on world- wide controls of Hg use and emissions remains elusive. Timely global actions on controls for Hg emissions, the use of Hg in consumer goods, and controls on artisanal uses of Hg, appear warranted.
v
TABLE OF CONTENTS
Title Page Declaration ...... ii Acknowledgements ...... iii Abstract ...... iv Table of contents ...... vi List of tables ...... xi List of figures ...... xiii Author’s note ...... xv
CHAPTER 1. INTRODUCTION 1.1 Mercury in the Global Environment ...... 1 1.1.1 Mercury Research in Remote Global Locations ...... 1 1.1.2 Mercury Emissions and the Global Atmospheric Burden ...... 3 1.1.3 Mercury and Human Health ...... 7 1.1.4 Mercury in Human History ...... 9 1.1.5 Transformations of Mercury in Environmental Media ...... 12 1.1.6 Mercury Research in Remote Oceania ...... 14 1.2 Literature Review ...... 15 1.2.1 Precipitation as an Indicator of Global Atmospheric Burden and Aerial Distribution of Hg ...... 17 1.2.2 Geologic Base as a Potential Source of Hg for Tutuila Aquatic Environments ...... 20 1.2.3 Bio-accumulation of Hg in Coral Reef Biota ...... 23 1.2.4 Risk Assessments and Consumption Limits for Reef Fish Fisheries ...... 24 1.3 Tutuila Island, Territory of American Samoa (USA) ...... 26 1.3.1 Location ...... 26 1.3.2 Geography ...... 27 1.3.3 Climate ...... 30 1.3.4 Geology and Petrography...... 32
vi
TABLE OF CONTENTS
Title Page CHAPTER 2. RESEARCH OBJECTIVES and STUDY PLANS 2.1 Atmospheric Wet Deposition of Hg in the Remote Southern Hemisphere ...... 34 2.1.1 Study Plan - Atmospheric Wet Deposition ...... 35 2.2 Distribution of Hg, Selected Elements, and Organic Carbon in Marine Sediments, Stream Suspended Solids, and Upland Soils in the Tutuila Environment ...... 37 2.2.1 Study Plan - Sediments, Stream Suspended Solids, and Soils ...... 40 2.3 Bio-accumulation of Hg in Tutuila Coral Reef Biota ...... 42 2.3.1 Study Plan - Bio-accumulation ...... 45 2.4 Human Health Risk Assessment for Hg in a Remote Artisanal Fishery ...... 49 2.4.1 Study Plan - Risk Assessment ...... 50
CHAPTER 3. MATERIALS and METHODS 3.1 Introduction ...... 51 3.2 Statistical Applications for Evaluating Hg in Environmental Matrices in Accordance with Research Objectives ...... 53 3.3 Study Sites and Sampling Stations; Un-impacted and Impacted Coral Reefs of Tutuila ...... 55 3.3.1 Masausi Bay; Masausi Catchment ...... 56 3.3.2 Alega Bay; Alega Catchment ...... 57 3.3.3 Loa; Pago Pago Catchment ...... 57 3.3.4 Amalau Bay; Vatia Catchment ...... 60 3.3.5 Tafeu Cove; Tafeu Catchment ...... 62 3.3.6 Fagafue Bay; Aasu Catchment ...... 63 3.4 Mercury in Rainfall and Marine Water ...... 65 3.4.1 Mercury in Rainfall and Monitoring Rainfall Events ...... 65 3.4.1.1 Rainfall sample collection ...... 65 3.4.1.2 Analytes, analytical methods, and detection limits for Hg in rainfall ...... 66 3.4.1.3 Blank correction for Hg in rainfall ...... 67
vii
TABLE OF CONTENTS
Title Page 3.4.1.4 Monitoring rainfall events ...... 68 3.4.1.5 Calculation of Hg wet deposition rates ...... 69 3.4.2 Mercury in Marine Water ...... 70 3.4.2.1 Marine water sample collection ...... 70 3.4.2.2 Analytes, analytical methods, and detection limits for Hg in marine water...... 71 3.4.2.3 Blank correction for Hg in marine water ...... 71 3.5 Mercury in Marine Sediments, Stream Suspended Solids, and Upland Soils ...... 72 3.5.1 Marine Sediments ...... 72 3.5.1.1 Marine sediments sample collection ...... 72 3.5.1.2 Additional elements in marine sediments ...... 73 3.5.1.3 Analytes, analytical methods, and detection limits for Hg and selected elements in marine sediments...... 73 3.5.1.4 Mineralogical analysis for marine sediments...... 76 3.5.1.5 Estimation of reduction/oxidation (redox) horizon in marine sediments ...... 76 3.5.2 Mercury in Stream Suspended Solids ...... 78 3.5.2.1 Stream water suspended solids sample collection ...... 78 3.5.2.2 Analytes, analytical methods, and detection limits for Hg and TOC in stream suspended solids ...... 79 3.5.2.3 Blank correction for Hg in stream suspended solids...... 79 3.5.3 Mercury in Upland Soils ...... 80 3.5.3.1 Upland soils sample collection ...... 80 3.5.3.2 Analytes, analytical methods, and detection limits for Hg in upland soils ...... 82 3.6 Mercury in Coral Reef Biota ...... 83 3.6.1 Turf Algae Collection ...... 83 3.6.2 Herbivorous Fish Collection (surgeonfish, Acanthurus lineatus) ...... 84 3.6.3 Carnivorous Fish Collection (goatfish, Mullidae spp.) ...... 85
viii
TABLE OF CONTENTS
Title Page 3.6.4 Carnivorous Fish Collection (barracuda, Sphyraena qenie) ...... 85 3.6.5 Analytes, Analytical Methods, and Detection Limits for Hg in Coral Reef Biota ...... 86 3.6.6 Calculation of Bio-accumulation Factors ...... 88 3.6.7 Assessment of Human Health Risks from Consumption of Tutuila Reef Fish ...... 89 3.6.7.1 Basis for risk-based consumption limits ...... 89 3.6.7.2 Calculation of risk-based consumption limits for Methyl Hg ...... 89
CHAPTER 4. RESULTS and DISCUSSION 4.1 Mercury in Tutuila Rainfall ...... 91 4.1.1 Acid Blank and Field Blank Correction for Rainfall Samples ...... 100 4.1.2 Paired Sampling Analysis for Rainfall Samples ...... 105 4.1.3 Rainfall for Period January 2007 through June 2008, Alega Village Station ...... 108 4.2 Mercury in Tutuila Marine Water ...... 111 4.3 Mercury in Tutuila Marine Sediments, Stream Suspended Solids, and Upland Soils ...... 118 4.3.1 Mercury in Tutuila Marine Sediments ...... 119 4.3.2 Total Organic Carbon in Tutuila Marine Sediments ...... 126 4.3.3 Mercury and Total Organic Carbon in Stream Suspended Solids ...... 129 4.3.4 Mercury in Tutuila and Aunu’u Upland Soils ...... 134 4.3.5 Selected Elements (Other than Hg) in Marine Sediments ...... 137 4.4 Mercury in Tutuila Coral Reef Biota ...... 139 4.4.1 Mercury in Turf Algae ...... 139 4.4.2 Mercury in Reef Fish ...... 144 4.4.2.1 Acanthurus lineatus ...... 144 4.4.2.2 Mullidae spp ...... 149
ix
TABLE OF CONTENTS
Title Page 4.4.2.3 Sphyraena qenie ...... 154 4.4.3 Bio-Accumulation Factors ...... 156 4.5 Human Health and Mercury in Tutuila Coral Reef Fish ...... 158 4.6 Hg in the Remote Tropical Environment of Tutuila ...... 163
CHAPTER 5. CONCLUSIONS and RECOMMENDATIONS 5.1 Conclusions ...... 168 5.2 Recommendations ...... 172
CHAPTER 6. REFERENCES ...... 174
APPENDIX A. LABORATORY QA/QC SUMMARIES ...... 214
APPENDIX B. REEF FISH SAMPLE LOGS ...... 237
APPENDIX C. ELEMENTS (other than Hg) IN MARINE SEDIMENTS ...... 245
x LIST OF TABLES
Table Page Table 3-1 Methods and detection limits for Hg in rainfall ...... 66 Table 3-2 Methods and detection limits for Hg in marine water ...... 71 Table 3-3 Methods and detection limits for Hg and elements in marine sediments ...... 74 Table 3-4 Methods and detection limits for Hg and TOC in stream suspended solids ...... 79 Table 3-5 Methods and detection limits for Hg in upland soils ...... 82 Table 3-6 Methods and detection limits for Hg in coral reef biota ...... 86 Table 4-1 Hg in Tutuila rainfall; volume-weighted mean ...... 95 Table 4-2 Estimated annual Hg wet deposition, Tutuila Island ...... 97 Table 4-3 Acid blank correction for HCl preservation of Tutuila rainfall samples ...... 101 Table 4-4 Field blank (FB) correction for Tutuila rainfall samples ...... 102 Table 4-5 Total Hg in Tutuila rainfall; acid and field blank corrected ...... 103 Table 4-6 Methyl Hg in Tutuila rainfall; field blank corrected ...... 104 Table 4-7 Relative percent difference (RPD) for Total Hg in paired samples for Tutuila rainfall ...... 105 Table 4-8 Relative percent difference (RPD) for Methyl Hg in paired samples for Tutuila rainfall ...... 106 Table 4-9 Monthly rainfall summary for Alega station, Tutuila Island ...... 109 Table 4-10 Measured rainfall for Alega station vs PRISM atlas prediction ...... 110 Table 4-11 Total Hg in Tutuila marine water ...... 112 Table 4-12 Methyl Hg in Tutuila marine water ...... 112 Table 4-13 TSS in Tutuila marine water ...... 117 Table 4-14 Total Hg in Tutuila marine sediments ...... 120 Table 4-15 Methyl Hg in Tutuila marine sediment ...... 121 Table 4-16 Redox horizon in vertical profile of Tutuila marine sediments ...... 123 Table 4-17 Total organic carbon in Tutuila marine sediment ...... 126 Table 4-18 DI water blank correction for Tutuila stream suspended solids ...... 130
xi LIST OF TABLES
Table Page Table 4-19 Total Hg in Tutuila stream suspended solids; blank corrected ...... 131 Table 4-20 Total organic carbon (TOC) in Tutuila stream suspended solids ...... 132 Table 4-21 Hg in Tutuila upland soils ...... 135 Table 4-22 Total Hg in Tutuila turf algae140 Table 4-23 Methyl Hg in Tutuila turf algae ...... 140 Table 4-24 Hg in Acanthurus lineatus muscle tissue ...... 145 Table 4-25 Hg in Parupeneus cyclostomus muscle tissue ...... 150 Table 4-26 Hg in muscle tissue for Sphyraena qenie ...... 155 Table 4-27 BAFs for Total Hg (referenced to marine water) ...... 157 Table 4-28 BAFs for Methyl Hg (referenced to marine water) ...... 157 Table 4-29 Consumption limits for MeHg in fish tissue ...... 159 Table 4-30 Consumption rates for un-impacted artisanal fishery ...... 161 Table 4-31 Consumption rates for impacted artisanal fishery ...... 162
xii LIST OF FIGURES
Figure Page Figure 1-1 Conceptual model of major components of the global Hg cycle ...... 5 Figure 1-2 Natural archive (ice core) of historic atmospheric Hg deposition; increasing background levels of Hg and influences of major volcanic events ...... 6 Figure 1-3 Major processes and Hg pathways in the biosphere ...... 8 Figure 1-4 Major chemical reactions of Hg transformations ...... 13 Figure 1-5 Study locus - Samoa Islands in remote Oceania ...... 27 Figure 1-6 Topography - Tutuila Island ...... 28 Figure 1-7 Major catchments - Tutuila Island ...... 29 Figure 1-8 Rainfall distribution - Tutuila Island ...... 31 Figure 3-1 Study sites - Tutuila Island ...... 55 Figure 3-2 Study site - Masausi (Masausi Bay) ...... 57 Figure 3-3 Study site - Alega (Alega Bay) ...... 58 Figure 3-4A Study site - Loa (Pago Pago Harbour)...... 59 Figure 3-4B Study site - Loa (Pago Pago Harbour) ...... 60 Figure 3-5A Study site - Vatia (Amalau Bay) ...... 61 Figure 3-5B Study site - Vatia (Amalau Bay) ...... 61 Figure 3-6 Study site - Tafeu (Tafeu Cove) ...... 62 Figure 3-7A Study site - Aasu (Fagafue Bay)...... 64 Figure 3-7B Study site - Aasu (Fagafue Bay) ...... 64 Figure 3-8 Upland soils sampling stations ...... 81 Figure 4-1 Total Hg in Tutuila rainfall (corrected) ...... 96 Figure 4-2 Methyl Hg in Tutuila rainfall (corrected) ...... 96 Figure 4-3 Total Hg in Tutuila rainfall (corrected) - paired sampling ...... 106 Figure 4-4 Methyl Hg in Tutuila rainfall (corrected) - paired sampling ...... 107 Figure 4-5 Alega station rainfall summary ...... 110 Figure 4-6 Rainfall distribution - Tutuila Island ...... 111 Figure 4-7 Hg in Tutuila marine water ...... 113 Figure 4-8 Hg in Tutuila marine sediments - all sites (mean w/SE) ...... 122 Figure 4-9 Hg in Tutuila marine sediment - un-impacted sites (mean w/SE) ...... 122
xiii LIST OF FIGURES
Figure Page Figure 4-10 Depth profiles for marine sediment sampling stations ...... 123 Figure 4-11 Total organic carbon in Tutuila marine sediment (mean w/SE) ...... 127 Figure 4-12 TOC vs Total Hg in Tutuila marine sediment - un-impacted sites ...... 128 Figure 4-13 TOC vs Methyl Hg in Tutuila marine sediment - un-impacted sites ...... 128 Figure 4-14 Total Hg in Tutuila stream suspended solids ...... 129 Figure 4-15 TOC in Tutuila stream suspended solids...... 130 Figure 4-16 Hg vs TOC in Tutuila stream suspended solids ...... 131 Figure 4-17 Total Hg and Methyl Hg in Tutuila turf algae (mean w/SE) ...... 141 Figure 4-18 Hg in muscle tissue - Acanthurus lineatus (mean w/SE) ...... 147 Figure 4-19 Fine sediment accumulation at base of turf algae (observed) ...... 147 Figure 4-20 Total Hg in muscle tissue vs weight - Acanthurus lineatus ...... 148 Figure 4-21 Methyl Hg in muscle tissue vs weight - Acanthurus lineatus ...... 149 Figure 4-22 Hg in Muscle tissue - Parupeneus cyclostomus (mean w/SE) ...... 151 Figure 4-23 Total Hg in muscle tissue vs weight - Parupeneus cyclostomus ...... 151 Figure 4-24 Methyl Hg in muscle tissue vs weight - Parupeneus cyclostomus ...... 152 Figure 4-25 Total Hg in muscle tissue vs weight - Mullidae spp ...... 153 Figure 4-26 Methyl Hg in muscle tissue vs weight - Mullidae spp ...... 153 Figure 4-27 Hg in muscle tissue - Sphyraena qenie ...... 155
xiv AUTHOR’S NOTE
The almost complete lack of experimental data for Hg in the Tutuila environment prior to this study precluded the traditional use of a priori hypothesis testing for scientific investigations of Hg on this remote oceanic island. Without a point of departure that was based on previous experimental data, the author was generally obliged to infer the relevance of Hg in the Tutuila environment, based on indicative, supportive, or suggestive evidence, that was experimentally-derived from a range of environmental matrices that were selected primarily to relate Hg occurrence on Tutuila to Hg in the global context. Investigative focus and field and analytical techniques used for this work were un-extraordinary. Mercury is a highly studied element, and research of similar scope and detail for Hg in environmental media has been conducted by many long-time practitioners in this field of interest. The unique value of this work derives from it being conducted on Tutuila Island. Repeated calls for increased research activities on Hg in remote global environments far from Hg emission sources have remained largely unanswered. Expense and operational difficulties for investigations of highly mobile Hg that occurs in trace concentrations in remote field conditions has hindered explorations in remote regions of the globe. On Tutuila, advantageous circumstances for funding and logistics were capitalised on to glean the maximum benefit from this opportunity to study Hg in this remote corner of Oceania. The results leave many more questions asked than answered. It is hoped that these outcomes will promote renewed interest in experimental research for Hg in remote global locations.
xv CHAPTER 1. INTRODUCTION
1.1 Mercury in the Global Environment
1.1.1 Mercury Research in Remote Global Locations
This thesis describes research on mercury (Hg) in environmental media from a remote part of the world that is far from any Hg emission source. Research was conducted to examine some of the prevailing notions of the distribution and environmental persistence of Hg at the global scale. Since the Hg poisonings at
Minamata and Niigata in Japan in the 1950s (McAlpine and Araki, 1958), there continues to be growing concern among professionals in the fields of environmental science and public health that Hg released to the environment from industrial activity and artisanal uses poses a contamination hazard for aquatic food webs on the global scale (e.g., Jernelöv and Lann, 1971; Nriagu, 1979; Fitzgerald et al., 1998; Poissant et al., 2002; UNEP, 2002; Pirrone and Mahaffey, 2006; UNIDO, 2007b). Evidence has accumulated that suggests aquatic systems in all regions of the globe are ultimately the recipients of anthropogenic releases of Hg, regardless of where discharges occur (e.g.,
Nriagu and Pacyna, 1988; Biester et al., 2002; Lamborg et al., 2002; Semkin et al.,
2005). The International Conference on Mercury as a Global Pollutant has convened eight times since 1990, yet, international consensus on the use of Hg and control of discharges to the environment remains elusive.
Mercury is a highly studied element; the body of literature is extensive, continues to grow, and presents consistent evidence for increased atmospheric concentration and deposition of Hg since the onset of industrialisation. Most Hg research, however, is from the industrialised regions of the northern hemisphere. The distribution and persistence of Hg in remote global locations remain largely un-described by direct experimental evidence, and our understanding of the global relevance of Hg pollution
1 therefore remains largely speculative. In the absence of robust pan-global experimental data from direct measurements of atmospheric concentrations of Hg and Hg in precipitation and other environmental compartments in remote regions, modeling was developed to evaluate the potential for long-range atmospheric dispersion and deposition of Hg (Petersen et al., 1996). Modeling continues to dominate as the principal technique used to support concerns for Hg as a global pollutant (e.g., Lamborg et al., 2002; Sunderland et al., 2008) and direct field investigations for Hg in remote regions of the world are rare.
Investigations for Hg in remote global locations are few because remote areas present major logistical and economic challenges for maintaining the quality of data sets. In areas that are not under the influence of local or regional emissions of Hg, concentrations of Hg in soil, sediment, and plant and animal tissues typically occur at sub-part-per-million levels, and at sub-part-per-billion levels in rainfall and surface waters. Protocols for field sampling, sample handling, and holding times, for volatile and highly mobile Hg species that occur at ultra-trace levels in environmental media are difficult to meet in remote global locations where facilities are unavailable and distances from analytical laboratories are great.
The research presented in this thesis answered a need expressed widely in the literature of the environmental science and public health fields; many more studies are needed that produce reliable data for the deposition and accumulation of Hg in ecosystems of remote global locations, especially in the tropics and southern hemisphere (e.g., Petersen et al., 1998; Bullock, 2000; Bullock and Brehme, 2002). The experimental results presented here for Hg deposition via rainfall, and the accumulation of Hg in marine biota on Tutuila Island in the remote South Pacific, were intended to enhance the literature on the occurrence of Hg in remote regions of the world that are
2 not under the influence from local or regional Hg emissions, and thus enhance our understanding of the potential for Hg as a global contaminant.
1.1.2 Mercury Emissions and the Global Atmospheric Burden
As the global integrator of nearly all Hg emissions (Figure 1-1), the atmosphere is the principal environmental reservoir from which Hg is ultimately deposited and made available for biological uptake and concentration in aquatic systems (e.g., Stein et al.,
1996; Boening, 2000; Grigal, 2002; Poissant et al., 2002). Contributions to the global atmospheric Hg reservoir include natural and anthropogenic sources.
Natural atmospheric inputs are geologic in origin, and are composed primarily of direct aerial emissions of Hg0. Direct (point-source) emissions are from volcanism and de-gassing of the Earth’s crust through fumaroles and at continental plate boundaries, and constitute the great majority of natural aerial emissions of Hg (Nriagu and Becker,
2003; Pyle and Mather, 2003). Principal indirect (non-point-source) natural inputs include Hg released from continental weathering and geothermal discharges at ocean ridges (Gustin, 2003). That Hg is a historical component of Earth’s atmosphere is demonstrated by long-term deposition records from lake sediments (Lamborg et al.,
2002; Fitzgerald et al., 2005; Sanders et al., 2008), peat bogs (Benoit et al., 1998;
Biester et al., 2002; Givelet et al., 2003) and glacial ice cores (Vandal et al., 1993;
Schuster et al., 2002).
Anthropogenic inputs of Hg to the atmosphere are industrial and artisanal in origin, and like natural emissions, are predominantly aerial emissions of Hg0.
In the industrialised countries, principal anthropogenic contributions to atmospheric
Hg are emissions from coal-fired power generation, municipal solid waste incinerators, medical waste incinerators, and smelting ores (US EPA, 1997; Pacyna and Pacyna,
3 2002; Pacyna et al., 2006). Other anthropogenic emissions of Hg in developed countries are non-point, and include various Hg compounds released to terrestrial and aquatic environments from waste burial, mining activities, and discharges from manufacturing processes, not all of which are sequestered in environmental matrices; an indeterminable fraction is eventually released to the atmosphere as Hg0 by transformation and evasion processes of the Hg cycle (US EPA, 1997; Hylander and
Meili, 2003). Estimates for non-point anthropogenic aerial emissions of Hg from industrialised countries are uncertain, but are thought to constitute a relatively minor fraction of total global anthropogenic Hg emissions to the atmosphere (US EPA, 1997;
Renner, 2005).
In developing countries, a largely unaccounted for source of non-point anthropogenic Hg emissions is the artisanal use of Hg for independent small-scale gold and silver mining, and precious metals recovery from waste electronic goods (e.g.,
UNIDO, 2007a, 2007b). Little data is available on the artisanal use of Hg world-wide, but it is expected to constitute a significant portion of the anthropogenically-derived releases of Hg to the atmosphere (Maxson, 2005). The magnitude of world-wide non- point source releases from artisanal uses of Hg has only been roughly estimated
(Maxson, 2005; UNIDO, 2007b).
Because of the extended residence time of Hg0 in the atmosphere (~1 yr) as a consequence of Hg0 volatility and chemical stability, there is extensive atmospheric mixing and dispersion, and the distribution of Hg in the global troposphere is generally described as near-uniform, except in regions under the influence of significant aerial discharges (Slemr et al., 1985; Lindqvist and Rodhe, 1985; Bergan et al., 1999; Poissant et al., 2002). There is evidence to suggest that there is a slight inter-hemispheric gradient of decreasing concentration from north to south (Fitzgerald, 1995; Slemr et al.,
4 1995; Lamborg et al., 1999, Temme et al., 2003), though measurements of atmospheric
Hg concentration from remote southern regions of the globe are few compared with data from northern temperate regions. The inter-hemispheric gradient is generally attributed to greater Hg emissions from industrial activity in the northern hemisphere compared with much lower emissions in the southern hemisphere.
Figure 1-1 Conceptual model of major components of the global Hg cycle (from Schroeder et al., 1989)
Current estimates are that natural and anthropogenic sources contribute about equally to the total global emissions of Hg (Nriagu, 1989; Gustin, 2003; Nriagu and
Becker, 2003; Pacyna et al., 2006). With a doubling of the total Hg emissions to the atmosphere since the onset of industrialisation, evidence indicates that the concentration of Hg in the global atmospheric reservoir has increased by a factor of 3-5 since the mid-
1800s (Slemr and Langer, 1992; Lamborg et al., 2002), which has consequently led to an increase of Hg deposition worldwide over the same period. Geologic records in natural archives of lake sediments, peat, and glacial ice (e.g., Figure 1-2) consistently indicate that the world-wide Hg deposition rate has increased substantially as a result of
5 the increased atmospheric concentrations of Hg from anthropogenic activities in the industrial age (Vandal et al., 1993; Benoit et al., 1998; Biester et al., 2002; Lamborg et al., 2002; Givelet et al., 2003; Fitzgerald et al., 2005; Sanders et al., 2008).
Figure 1-2 Natural archive (ice core) of historic atmospheric Hg deposition; increasing background levels of Hg and influences of major volcanic events (from Schuster et al., 2002)
The profile in Figure 1-2, from two cores in the Upper Fremont Glacier (Wyoming,
USA), clearly indicates significant natural and anthropogenic events for aerial releases of Hg over the past three centuries, and an overall increase in atmospheric deposition concomitant with the industrial age.
6 1.1.3 Mercury and Human Health
Once Hg is deposited to terrestrial and aquatic environments, the Hg compounds formed in abiotic and biotic transformation pathways are of concern because of their environmental mobility, propensity to bio-accumulate, and neurotoxicity. Once Hg is deposited, biological uptake can result in bio-magnification of Hg compounds by more than a million-fold between abiotic and biological compartments. Of particular importance are the alkylated Hg compounds, and of these, methylmercury is of greatest concern for human health (e.g., Dales, 1972; Chang, 1977; ATSDR, 1999; Yokel et al.,
2006). Methylmercury (MeHg) is formed primarily from inorganic Hg(II) species by the metabolic processes of microbes in soils, water, and sediments (Figure 1-3) (e.g.,
Wood et al., 1968; Jensen and Jernelöv, 1969; Kitamura et al., 1971; Landner, 1971;
Vonk and Sijpesteijn, 1973; Gilmour et al., 1998, Fleming et al., 2006).
Bio-accumulation of MeHg is enhanced in aquatic food webs compared with terrestrial systems (Harriss, 1971; Jernelöv and Lann, 1971; Nriagu, 1979; ATSDR,
1999). As a strong bio-accumulative, Hg in the aquatic environment can result in a 106-
107 increase in Hg concentrations in tissues of aquatic organisms (Lindqvist, 1991;
Mason et al., 1995). Methylmercury predominates in muscle tissue of most top predator fish and typically comprises 95-99% of total Hg (THg) (Bloom, 1992). Ingestion of
Hg-contaminated fish and shellfish is the primary exposure pathway of concern for the general population (US EPA, 1997; ATSDR, 1999; US EPA, 2000).
7 Figure 1-3 Major processes and Hg pathways in the biosphere (from US Geological Survey, 1996)
Note: DOM = dissolved organic matter
Mercury is non-essential in human nutrition, and is not known to be part of the metabolism in any species (ATSDR, 1999). Methylated Hg compounds readily cross the blood-brain barrier and the placental membranes in mammals (Chang, 1977; US
ATSDR, 1999; Castoldi et al., 2001). Methylated Hg is particularly dangerous to the developing foetus and can result in impaired neurological development (Marsh et al.,
1987; Marsh et al., 1995; Murata et al., 1999; Weihe et al., 2002; Jedrychowski et al.,
2007). Infants can be exposed to MeHg through breast-milk (ATSDR, 1999). In children and adults, exposure to Hg can result in parathesia, impaired cognitive function, impaired motor skills, renal failure, and convulsions (Myers et al., 2000).
Extreme exposure leads to permanent cognitive impairment, physical disability, or death, as shown by the incidences at Minamata and Niigata in Japan (D’Itri and D’Itri,
1977; Tsubaki and Irukayama, 1977).
8 With a doubling of total world-wide Hg emissions since the pre-industrial era and the mounting evidence for the proliferation of Hg globally via atmospheric processes, the implications for human health are manifest. An expected consequence from industrial-era discharges of Hg and extensive dispersion is a world-wide increase in the burden of Hg residues in tissues of aquatic organisms, and increased world-wide potential for adverse human health impacts from consumption of aquatic species.
1.1.4 Mercury in Human History
Mercury in human affairs reaches back to around 2000 B.C (Farber, 1952). The chemical symbol “Hg” is derived from the Latin hydrargyrus, a combinational form for
“water-silver”. Mercury production from cinnabar ore (HgS) is simple and efficient, and ancient peoples could readily process essentially pure metallic Hg for use in medicines, ointments, cosmetics, and cultural ceremonies. Mercury is found in ancient tombs from China, Egypt, and India, and was used in decorative displays in royal courts and for religious ceremonies.
The ability of Hg to amalgamate with precious metals was recognized as early as
500 B.C (D’Itri and D’Itri, 1977), and alchemists and gold and silver miners created a steadily growing demand for Hg over the next two millennia. Other than mining, among the greatest pre-industrial demands for Hg was the preparation of mercuric nitrate for felting hats in Europe, America, and the European colonies in the eighteenth and nineteenth centuries. Severe neurological disorders were observed as common ailments among the unfortunate hatters, and were the first indications of dangers from occupational exposure to Hg (D’Itri and D’Itri, 1977).
With the onset of the industrial era, the chemical and physical properties of Hg led to greatly expanded uses in more than 3000 manufactured goods and industrial processes
9 by the middle of the twentieth century (D’Itri and D’Itri, 1977; US EPA, 1997; US
EPA, 2008). Mercury is a component in many applications for instrumentation, control systems, lighting, telecommunications, and other electronics. As a catalyst, Hg is important in manufacturing processes for petrochemicals, plastics, and many intermediate compounds for the chemical industries. Once used widely as a fungicide, this use is curtailed in many developed countries, but is thought to continue un- regulated and un-accountably in many parts of the world. Until recently, Hg was used almost exclusively as the flowing cathode in the chlor-alkali cell process that produces
Cl2 gas and NaOH from seawater. Beginning in the early twentieth century, discharges of Hg to estuarine systems from chlor-alkali facilities were particularly egregious, until many industrial nations implemented alternative processes by the 1980s.
Although Hg had earlier been recognized as a potential hazard to human health, discharges to land, water, and air were not acknowledged as a significant concern until the late 1950s, when illnesses and deaths among thousands of Japanese villagers were attributed to poisoned fish and shellfish as a result of industrial discharges of Hg to
Minamata Bay, and later at Niigata (McAlpine and Araki, 1958; D’Itri and D’Itri,
1977). Investigations to determine the cause of “Minamata disease”, a condition manifested by slurred speech, ataxic gait, cognitive impairment, birth abnormalities, and sometimes death, revealed that inorganic Hg undergoes transformation in the biosphere to highly neurotoxic methylated forms, and aquatic organisms at upper trophic levels were subsequently found to be particularly vulnerable to bio-accumulation and bio- magnification of methylated Hg species (Jernelöv and Lann, 1971; Lindqvist, 1991;
Mason et al., 1995). The legacy of the Japanese poisonings has spanned two generations
(Sakamoto et al., 1991; Takaoka, et al., 2008).
10 As a result of increased awareness after the incidents of Minamata and Niigata, some industrialised nations began efforts for controls on Hg use and emissions as early as the
1970s. Today, regulatory controls on the use of Hg in industrial processes and manufactured goods, and the discharge of Hg wastes, are relatively advanced in North
America and Europe, though controls on point-source atmospheric discharges lag far behind controls on discharges to terrestrial and aquatic environments (Pirrone and
Mahaffey, 2006; Selin and Selin, 2006; Swain et al., 2007).
World-wide production of Hg peaked at ~10,000 tonnes annually in the 1960s, declining to ~3500 tonnes annually in 2005 (Maxson, 2005). Today, Hg is an internationally traded commodity (9000 tonnes annually in 2000) with a net flow from industrialised to developing countries (Swain et al., 2007; UNIDO, 2007a, 2007b).
There are now some 12,000 tonnes of metallic Hg stock-piled in Europe and North
America from decommissioned chlor-alkali facilities, and it is thought that much of this
Hg is transported un-accountably across international boundaries to the developing world and supports the small-scale artisanal uses for mining or recovery of metals from electronics wastes (Swain et al., 2007, UNIDO, 2007a). Only recently have the
European Union and the United States proposed legislation to ban exports of Hg, to take effect in 2011 and 2013, respectively (UNIDO, 2007a; Van Noorden, 2008).
Multi-scale governance at the local, regional and global scales is proposed as the only practical option for effective management of global Hg proliferation, because of the complexity of trade routes and the obscurity of transactions in the Hg market (Selin and Selin, 2006; Swain et al., 2007; UNIDO, 2007a).
As presented in this thesis, the development of experimentally derived data on the occurrence and persistence of Hg in remote global locations, to quantify the environmental burden of Hg and the potential for human health and environmental
11 impacts, may help enhance policy- and decision-makers’ efforts for greater international controls on Hg production and use.
1.1.5 Transformations of Mercury in Environmental Media
Mercury occurs in three oxidation states in environmental media, Hg0, Hg(I), Hg(II), of which only Hg0 and Hg(II) are important in the environmental chemistry of Hg.
Oxidation state determines Hg volatility and reactivity, and thus the likelihood for Hg dispersion, deposition, and sequestration (Nriagu, 1979; Bodek et al., 1988; Lindberg and Stratton, 1998; Lin and Pehkonen, 1999; Yokel et al., 2006).
Elemental Hg0 is the predominant species in atmospheric Hg, and because of its volatility and stability (Nriagu, 1979; Bodek et al., 1988), Hg0 is transported long distances in the atmosphere and may reside there for as long as two years before it is deposited and sequestered in abiotic or biological media (Lindqvist and Rodhe, 1985;
Slemr et al., 1985). Atmospheric Hg0 is part of a complex flux of deposition and re- emission, where many intermediate compounds are formed along abiotic and biotic pathways through a series of oxidation and reduction reactions before Hg is returned to the atmosphere as Hg0 (e.g., Stein et al., 1996; Lin and Pehkonen, 1999; Pirrone and
Mahaffey, 2006). Major chemical transformations for Hg in the atmosphere include:
0 ● gaseous phase oxidation of Hg to Hg(II) by O3, Cl2 and NO3 (“●” denotes radical); aqueous phase oxidation of Hg0 to Hg(II) by ●OH and OCl-; and aqueous phase
0 ● 2- reduction of Hg(II) to Hg by HO2 and SO3 (Nriagu, 1979). By the major chemical transformations (Figure 1-4) the atmospheric reservoir is continually gleaned of Hg
(Nriagu, 1979; Stein et al., 1996).
Experimental data (Nriagu, 1979) show that of the three oxidation states, Hg0 and
Hg(II) are the species of principal interest for environmental transformations of Hg, and
12 that Hg(I) is of little importance. In the gaseous phase, Hg(I) is thermodynamically improbable because common oxidants in the atmosphere will oxidize Hg0 directly to
Hg(II). The reduction of Hg(II) to Hg(I) is similarly improbable because Hg(I) will be quickly reduced to Hg0 by common atmospheric reductants.
Figure 1-4 Major chemical reactions of Hg transformations (from Stein et al., 1996)
Note: DOC = dissolved organic carbon
In the aqueous phase (cloudwater and raindrops) Hg(I) occurs as Hg2(II), which is extremely unstable in the presence of atmospheric ligands, and will disproportionate rapidly into Hg(II) and Hg0. Given the instability of Hg(I), it will occur only as a rare and transitional species and does not play a significant role in the chemical transformation of atmospheric Hg to more complex and environmentally important compounds.
13 The most important species for atmospheric deposition and biotic uptake is Hg(II), as shown in Figures 1-3 and 1-4, because of its solubility in aqueous media and the ability to adsorb onto atmospheric particles, which increases the propensity for deposition via precipitation (Nriagu, 1979; Bodek et al., 1988; Lin and Pehkonen,
1999).
1.1.6 Mercury Research in Remote Oceania
Adherents to the principle of Hg as a global pollutant cite elevated Hg levels found in marine flora and fauna from remote locations as an indication of significant Hg inputs to remote aquatic environments, and assert that these levels cannot be accounted for by natural sources alone (e.g., Fitzgerald et al., 1998). Findings of increased deposition of Hg in the Arctic (Fitzgerald et al., 2005; Steinnes and Sjøbakk, 2005), and elevated Hg in humans and biota from remote and unpolluted areas such as Greenland
(Johansen et al., 2007; Rigét et al., 2007) support concerns for the global proliferation of Hg, and consequential environmental and human health impacts world-wide.
If anthropogenic releases of Hg are shown to increase deposition and environmental accumulation of Hg in regions that are far from emission sources, there are important trans-boundary implications for emitters, and all anthropogenically-derived Hg emissions are therefore of global concern, and not solely that of local, regional, or national jurisdictions. For remote areas of the world where economic or dietary subsistence depend on the continued viability of aquatic species, the potential globalisation of Hg presents important questions of trans-boundary imposition of risk, accountability, and environmental justice (O’Neill, 2004).
Mercury as a global contaminant is an issue that demands timely resolution. The world’s most rapidly expanding economies (Asia region) now produce more than half
14 the global anthropogenic atmospheric Hg, and total emissions from these sources are expected to increase substantially by 2025 (Renner, 2005; Wong et al., 2006). The expanding Asian economies have embraced abundant and relatively inexpensive coal as an alternative to petroleum or other fuels, and are rapidly expanding the industrial uses of coal for steel-making and ore smelting. These economies also exhibit an ever increasing demand for Hg for use in artisanal gold and silver mining, and for precious metals recovery from waste electronic goods (Wong et al., 2006). Similar to the events of Minamata and Niigata half a century ago, it appears that national policy- and decision-makers continue to take a conservative approach to the regulation and control of Hg emissions, in favour of economic expansion.
Contemporary concerns for Hg emissions are by no means attributable only to the
Asia region. Although the heavily industrialised economies of Japan, Europe, and
North America have achieved some substantial reductions of Hg emissions in recent decades (Pacyna et al., 2006), emissions from these sources still constitute nearly half the total anthropogenic global emissions of Hg.
International consensus and action on initiatives to control Hg emissions world-wide seem unlikely without reliable and defensible data to improve our understanding of Hg as a global pollutant. The research presented for this thesis, which examines the deposition and accumulation of Hg in the environment of Tutuila Island in the remote
South Pacific, far from any Hg emissions, is a contribution towards an international resolution on the global relevance of Hg pollution.
1.2 Literature Review
A literature review was conducted to highlight some important knowledge gaps for the occurrence of Hg in remote “un-impacted” environments, and to serve as the basis
15 for specific research objectives (Chapter 2). The literature review focused on Hg deposition, the potential for natural sources of Hg to confound studies on the impacts of anthropogenically-derived emissions of Hg in remote areas, and bioaccumulation and health risks from seafood consumption in aquatic environments that are far from any Hg emissions.
Based on published sources, research on Hg appears to be heavily biased towards
North America and Europe. No published sources were found for Hg in environmental media from the Middle East or Africa, and no contemporary sources were found for
India, or Russia. It is possible that research on Hg in these regions is published but not widely circulated, and not included in the English language literature or abstracting systems.
In the un-published literature, information on Hg in environmental media may be substantial, and may compensate somewhat for the bias that is apparent in the published realm. Reports and studies, generated by industry and manufacturing for compliance purposes, or by local or regional public health authorities regarding exposure to Hg, or by regulatory authorities for determinations for remediation actions, are undoubtedly numerous. At the international scale, the amount of unpublished information on Hg remains indeterminable, but may well exceed what exists in the published literature. A survey of the literature for the purpose of this thesis could not purport to cover such a range of material, and with few exceptions, only peer-reviewed published literature were included in this review.
16 1.2.1 Precipitation as an Indicator of Global Atmospheric Burden and Aerial Distribution of Hg
Published material from the past 30 years revealed that Hg in precipitation is very well documented for much of North America and Europe, much less so for the Arctic, and minimally for parts of Eastern Asia.
World-wide, North America currently has the greatest coverage for monitoring Hg in precipitation. The Mercury Deposition Network (MDN) as part of the National
Atmospheric Deposition Program (NADP), initiated in 1996, currently operates ~100 active monitoring stations to monitor Hg in precipitation in Alaska, Canada, Mexico, and the co-terminous USA, with a further 30 inactive stations in place for programme redundancy or expansion. The MDN stations are operated through ~250 cooperative agreements between USA state and federal scientific authorities of local or regional jurisdictions, and commercial laboratories, under the auspices of the US Department of
Agriculture (USDA, 2007). Other significant Hg monitoring programmes conducted in
North America include: the Florida Atmospheric Mercury Study (FAMS), which operated ~10 stations to monitor Hg in precipitation in Florida for a period of 5 years
(Guentzel et al., 1995; Guentzel et al., 2001); a 2-year, 3-site study in Michigan (Hoyer et al., 1995); a 4-year multi-site study in the Chesapeake Bay region (Mason et al.,
2000); a multi-site programme operating since 2002 in eastern Ohio (Keeler et al.,
2006); a 2½-year study recently completed at a single station in upper New York State
(Lai et al., 2007); and a 2-year single station study at the Experimental Lakes Area in northwestern Ontario, Canada (St. Louis et al., 1995). Among the North America studies, Arctic locations are least numerous, and often involve snowfall rather than rainfall, though both are investigated when possible (e.g., Semkin et al., 2005; Poulain et al., 2007).
17 The review of published sources for research on Hg in North America precipitation was not intended to be exhaustive, and there are numerous studies other than those mentioned here that document Hg in North America precipitation, often collaterally as part of broader research objectives.
Compared with North America, there is far less information on Hg atmospheric deposition for the remainder of the world. For South America, only a few studies were found that presented experimental data for Hg in precipitation (e.g., Fadini and Jardim,
2001; Jardim and da Silva, 2003), none of which were for remote areas, with most Hg- precipitation studies for this continent focused on the use of predictive modeling. In
Europe, investigations of Hg in precipitation are generally concentrated among the
Nordic and Baltic states (e.g., Iverfeldt, 1991; Munthe et al., 1995; Leermakers et al.,
1997; Larssen et al., 2008). In Asia, China has exhibited an expanded interest in Hg in environmental media in recent years, but published sources found were limited to atmospheric concentration of Hg (e.g., Zhang and Wong, 2007), and gaseous or aqueous industrial discharges related to impacts on surface waters and soils, and Hg related to coal (e.g., Wang et al., 2006; Zheng et al., 2007). In Japan, most research on
Hg has heretofore focused on human health implications, much of which does not appear in the English language and does not receive wide circulation. Only recently have contributions on Hg in precipitation from this important geographic location of the industrialised northern hemisphere been made available in the English language literature (e.g., Sakata and Marumoto, 2005; Sakata et al., 2006; Sakata and Asakura,
2007).
Past studies in remote locations in northern and southern hemispheres suggest the globalisation of Hg pollution via atmospheric processes, but studies are few, global coverage is sparse, and sampling limited. For remote areas of the northern hemisphere,
18 Fitzgerald (1989) reported 2.8 ±1.6 ng L-1 THg (n=3) from Enewetak Atoll (11o N, 165o
E), and 9 ±5 ng L-1 THg (n=5) from the North Pacific between Hawaii and Alaska (44o-
55o N, 160o W). Bloom and Watras (1989) reported 2.80 ±1.44 ng L-1 THg and 0.139
±0.061 MeHg (n=10) on the Olympic Peninsula of Washington State, USA (48o N, 125o
W). No other published sources were found for Hg in precipitation in remote northern hemisphere locations.
For the remote southern hemisphere, published information for Hg in rainfall is limited to just four locations. Gill and Fitzgerald (1987) reported 3.8 ±0.8 ng L-1 THg
(n=6) from the Tasman Sea (35o S, 170o E), and 4.4 ±2.4 ng L-1 THg (n=4) from
American Samoa. Mason et al. (1992) reported THg of 2.88 ±1.3 ng L-1 (n=5) in the equatorial mid-Pacific between Samoa and Tahiti. In the equatorial South Atlantic,
Lamborg et al. (1999) reported THg of 3.33 ±1.7 ng L-1 (n=5).
Limited historical data for Pacific Ocean sectors of the northern and southern hemispheres suggest little difference in Hg concentration in rainfall among inter- hemispheric locations in the recent past (~20 years), which supports prevailing thought on the globalisation of Hg. Although atmospheric concentration of total gaseous Hg
(TGM) has shown a slight gradient in the Pacific hemisphere from the north (~2 ng m-3) to the south (~1 ng m-3) between 55o N and 55o S latitudes (Fitzgerald, 1995), the inter- hemispheric gradient of TGM does not appear to be reflected significantly in historical data for rainfall Hg concentrations among Pacific inter-hemispheric locations.
Before this study it was not possible to make conclusive comparisons for Hg in precipitation among inter-hemispheric locations. Although there are currently several
MDN locations in North America that have suitable remote location characteristics to serve as remote global background sites for Hg in precipitation, and from which reliable data is available, there are no known contemporary studies for the remote southern
19 hemisphere to make comparisons with. No published data since Lamborg et al. (1999) was found for Hg measured in rainfall in the southern hemisphere. Research objectives for Hg in Tutuila Island rainfall (Chapter 2) were intended to fill the data gap for Hg in southern hemisphere precipitation.
1.2.2 Geologic Base as a Potential Source of Hg for Tutuila Aquatic Environments
In the absence of direct discharges of Hg, it is widely held that atmospheric deposition is the most important source of Hg in aquatic biota (e.g., Fitzgerald et al.,
1991; Fitzgerald et al., 1998; Schroeder and Munthe, 1998). Conclusive determinations for the globalisation of Hg, developed from studies in remote global locations, however, require that natural sources of Hg in remote environments should be accounted for
(Rasmussen, 1994; Downs et al., 1998). The geologic base of Tutuila is volcanic in origin (see Section 1.3.4) and is composed primarily of olivine basalts (MacDonald,
1944; Stearns, 1944). Parent material is the original source of naturally occurring Hg in soils, and since basalts can be enriched with Hg, basalt parent material can lead to natural enrichment of soil Hg (Andersson, 1979).
In the case of Tutuila Island, no data was found in the published or un-published literature for investigations of Hg in parent rocks, so the influence of Tutuila’s volcanic origins on present-day soil Hg remains indeterminable. The period of geologic and petrographic investigations for Tutuila preceded the period of heightened awareness for concerns of Hg in the environment, so Hg might have seemed irrelevant to past geologic investigators. Moreover, trace-elements field and analytical techniques are time-consuming and costly, and since Hg does not correlate significantly with other elements, it is generally of little interest in conventional mineralogy (Carlson, 2005).
20 Since there is no data on Hg in Tutuila rocks, some inferences were made for the likelihood of a significant geologic contribution of Hg from native basalts to Tutuila soils. Inferences were based on Hg in basalts from other locations, and the experimental data for Hg in Tutuila soils from this study (see Chapter 4).
Specific information for Hg in basalts is sparse in the published scientific literature.
Among the most recent reviews of available data, it is indicated that Hg in basalts is variable, generally in the range of 0.002-0.035 mg kg-1 (Carlson, 2005). Turekian and
Wedepohl (1961) prepared a compilation of sources for the distribution of the elements in the Earth’s crust that included Hg data from composites of German rocks and reported mean Hg of 0.09 mg kg-1 in basalts. Gladkikh et al. (1975), from their extensive research on Hg in volcanic rocks from the Soviet Union, which included more than 400 rock samples, showed that Hg in olivine basalts was generally ≤ 0.010 mg kg-
1, with basalts of higher silica content having generally lower Hg than those containing less silica. Mercury incorporation in silicate mineralisation (basalts) is limited by the relatively large size of the Hg2+ ion, and its relatively low ionic activity, and concentrations of around 0.010 mg kg-1 Hg in basalts and other igneous rocks are typical (Geological Survey of Finland, 2007). Flanagan et al. (1982) measured Hg in basaltic reference materials and found mean Hg of 0.007-0.034 mg kg-1. Flanagan et al. found that Hg content in basalts of island arcs and the circum-Pacific region had less Hg than similar basalts from continental areas, with concentrations generally at the lower end of the reported range. The Flanagan data includes basalt from the Kilauea Volcano
(Hawaii Island, USA), with Hg of 0.007 ±0.0005 mg kg-1, which was the most appropriate available data found in the literature for direct comparison with Tutuila basalt. Values for Hg in basalts similar to Flanagan et al. were found by Cameron and
Jonasson (1972), who analysed 98 samples of volcanic basalts, andesites, and rhyolites,
21 and found mean Hg of 0.0084 mg kg-1. Wedepohl (1995) reported a range of ~0.040 mg kg-1 for Hg in basalts, mostly from unpublished sources. Recent data from the tectonically active California coast (USA) shows Hg in basaltic continental rocks of
0.017-0.028 mg kg-1 (Smith et al., 2008), although one anomalously high value of 0.288 mg kg-1 was found in this region of potential enrichment and mobility of Hg.
Summarising the information found for Hg in basalts, indications are that Hg occurs in basalts in the range of ≤0.010-0.040 mg kg-1. Based on the frequency of values and sources for this range, it can be reasonably inferred that Tutuila basalts, predominantly magnesium-iron silicates of oceanic origin, contain Hg at the lower end of the range.
Lower range values for Hg in Tutuila basalts could also be supported by the value reported for Hg in Kilauea basalt, though the basis of this comparison - similarity as
Pacific plate hotspot shield volcanos - has recognized limitations.
In the absence of an external source of Hg to the Tutuila environment, weathering processes would be expected to attenuate the Hg concentration in Tutuila soils compared with Hg in the island’s parent material (Ure and Berrow, 1982). Thus, in the absence of an external Hg input to Tutuila, Hg in Tutuila soils could reasonably be expected to be at, or less than, the lower range of values given for Hg in basalts
(≤0.010-0.040 mg kg-1).
Research objectives for determining the range of Hg in Tutuila soils, as presented in
Chapter 2, were developed to evaluate the influence of Hg atmospheric deposition on soil Hg, and to evaluate the likelihood of a significant geologic source of Hg in
Tutuila’s aquatic environment.
22 1.2.3 Bio-accumulation of Hg in Coral Reef Biota
An important factor in determining the relevance of Hg as a global pollutant is the occurrence and concentration of Hg in tissues of aquatic biota in remote regions of the globe that are far from any Hg emission sources. Aquatic biota can be sensitive indicators of Hg pollution because of the strong bio-accumulative characteristics of Hg in the aquatic environment (e.g., Westöö, 1966; Jernelöv and Lann, 1971; Nriagu, 1979;
Downs et al., 1998). Accumulation factors of 105-107 between water and upper trophic biota are commonly found for Hg in lacustrine, riverine, and estuarine environments
(Lindqvist, 1991; Watras and Bloom, 1992; Mason et al., 1995). Although there is compelling evidence for the global atmospheric distribution of Hg, and an increase in
Hg deposition rates world-wide since the onset of industrialisation, geographical coverage by studies on Hg occurrence and accumulation in aquatic biota from which to evaluate global trends and patterns is markedly deficient.
Studies on bio-accumulation of Hg in aquatic flora and fauna are most extensive for freshwater systems, typically with relevance to important food and economic resources of the northern hemisphere, especially for North America and Europe. Based on the number of publications, there has been increased interest in research on Hg in marine biota during the past ~10-15 years, but this body of literature is still far less extensive than that for Hg in biota of freshwater systems. In remote global locations studies on bioaccumulation in marine systems are rare, and it remains inconclusive to what extent the Hg increase in the atmospheric reservoir due to anthropogenic emissions is reflected in the levels of Hg in aquatic biota in remote and seemingly un-impacted environments.
For coral reefs, no information was found for studies on Hg in biota for reefs not under the influence of Hg emissions. Several important studies on Hg in Pacific tropical reef biota are included in the literature (Denton et al., 2006; Chouvelon et al.,
23 2009; Denton and Morrison, 2009), but study locations were not representative of remote un-impacted sites.
Besides a general lack of adequate geographic representation for research on bio- accumulation in remote un-impacted marine systems, two other factors limit the synthesis of meaningful conclusions about patterns or trends in Hg accumulation in biota world-wide. These are differences in species studied, and differences in habitats, among diverse global study locations. There were no studies found in the literature for remote un-impacted locations that were suitably consistent in species or habitat parameters to allow for a meaningful evaluation of Hg bio-accumulation at spatial or temporal scales. It is therefore currently not possible to synthesise the extensive body of literature on Hg bio-accumulation to formulate conclusions on Hg as a global pollutant.
At this time, studies on Hg bio-accumulation in remote global locations far from Hg emission sources must be interpreted individually as indications of the potential for Hg as a global pollutant. Data on bio-accumulation of Hg in Tutuila Island reef biota, as presented in this thesis, is similarly restricted in this regard.
1.2.4 Risk Assessments and Consumption Limits for Reef Fish Fisheries
Risk assessments for exposure to Hg through seafood consumption render a practical synthesis of environmental data with human and environmental health concerns
(localised), from which to draw inferences for the occurrence and proliferation of Hg in the global environment.
Since the poisonings at Minamata and Niigata, consumption of aquatic species is firmly established as the principal pathway of exposure to Hg in the general human population (Irukayama et al., 1962; Jernelöv and Lann, 1971; Mason et al., 1995; US
24 EPA, 1997; ATSDR, 1999; Mergler et al., 2007). According to interests popularly expressed throughout the relevant literature, human health concerns appear to remain the primary motivation for studying Hg in the global environment. Environmental
(ecological) health is historically of lesser concern, except as it relates to human health, as demonstrated by the anthropocentric context that dominates current and past research on Hg in the environment.
Similar to most published research on Hg in the environment, human health risk assessments are almost exclusively limited to the developed nations of the northern hemisphere. As an example, as of 2005, 44 of 50 states in the USA had issued fish advisories for inland and coastal waters. Compared to the USA, there appears to be a less conservative approach to risk assessments among other developed countries of the northern hemisphere, notably Canada, Japan, Taiwan and members of the EU, and similarly in Australia in the southern hemisphere, and although overall there are fewer studies in the published and unpublished literature for these nations than for the USA, studies are still numerous.
There are few human health risk assessments for seafood consumption completed for remote global locations, and studies on the potential impacts on human health from exposure to Hg from seafood in remote artisanal fisheries are virtually unknown. For the remote tropical Pacific, only two sources for risk assessments for artisanal fisheries were found. In American Samoa, several screening-level studies were completed between 1990 and 1994 (AECOS, 1991; US EPA, 1992; EnviroSearch, 1994), and following on these, one detailed risk assessment was completed (Peshut and Brooks,
2005; Peshut et al., 2008). In New Caledonia, Chouvelon et al. (2009) analysed THg and MeHg in tissues from a wide variety of crustaceans, bivalves, cephalopods, and fish. However, risk assessments by Peshut and Brooks, Chouvelon et al., and Peshut
25 could not be interpreted as representative of un-impacted remote global locations because of the proximity of human habitation to study sites. In American Samoa, study sites were mainly along the populated south shore of the island, and within Pago Pago
Harbour. All collection sites for the New Caledonia study were located within the main lagoon, and hence subject to anthropogenic influences from waste discharges related to shore-side development and mining activities in the uplands.
Examination of the published literature with regard to the global relevance of Hg pollution highlighted significant gaps in our knowledge for human health risks from exposure to Hg in artisanal fisheries of remote, un-impacted environments. Specific research objectives to address these knowledge gaps are presented in Chapter 2.
1.3 Tutuila Island, Territory of American Samoa (USA)
1.3.1 Location
American Samoa is comprised of five volcanic high islands and two coral atolls
(total land area 195 km2) located in the eastern half of the Samoa Islands archipelago, in the region of 11o-14o south latitude, 168o-171o west longitude, in the South Pacific
Ocean (Figure 1-5). Tutuila Island is the largest (137 km2) and most populous
(~56,000) island in the group, and is the centre of Territorial government and commerce
(US Department of Commerce, 2001). The location for Tutuila Island is typically specified as the main port facilities in Pago Pago Harbour at 14o 17' south latitude, 170o
41' west longitude.
An appreciation of Tutuila’s remote global position is given by reference to neighboring world cities (great circle routes, in statute km). From Pago Pago,
Honolulu, Hawaii lies 4200 km to the north-northeast; San Francisco, California lies
7700 km northeast; and, Auckland, New Zealand lies 2900 km to the south-southwest.
26 To the west is Suva, Fiji at 1250 km, and to the southwest is Sydney, Australia at 4700 km. Other distant points of reference are Easter Island at 6500 km to the east-southeast, and Lima, Peru at 8900 km to the east.
Landmasses between Tutuila and given points of reference are limited to small oceanic islands, most of which are uninhabited or have small subsistence populations, amid large expanses of open ocean. Because of its extreme isolation from up-current atmospheric or oceanic point-source emissions of Hg, Tutuila Island is well-situated as a remote study location for investigating the global relevance of Hg pollution.
Figure 1-5 Study locus - Samoa Islands in remote Oceania (from GIS Users Group, American Samoa Government, 2006)
1.3.2 Geography
All of the American Samoa islands are small, and, except for Tutuila, sparsely populated or uninhabited. Topography of Tutuila is dominated by steep slopes with dense tropical forests, shallow soils, frequent rock outcroppings, and numerous small
27 and sharply defined catchments (Figures 1-6 and 1-7). Except as otherwise noted, geographic and population parameters summarised for Tutuila are taken from A Coastal
Zone Management Atlas of American Samoa (Wingert, 1981).
Figure 1-6 Topography - Tutuila Island (study sites indicated) (from GIS Users Group, American Samoa Government, 2006)
Tutuila Island is oriented east-west along its major axis, extending 32 km from Cape
Matatula on the east to Cape Taputapu on the west, and <10 km along the north-south minor axis. The coastline of Tutuila (~200 km) is extremely irregular with numerous small bays open to the sea, which lacking bars or barrier reefs, are subject to the sea conditions of exposed coasts. Most of the coastline is rocky with abrupt elevation changes immediately above breaking waves. Beaches are few, and are typically short
28 stretches of mostly coralline sand mixed with some weathered basalts, in the inmost areas of the bays in which they occur.
A sharp-crested ridge extends the entire east-west length of the island and effectively divides Tutuila into northern and southern regions. Tutuila has no estuaries and no lacustrine waters. Surface waters are limited to a few dozen perennial streams, most of which have short steep reaches, and typically low base flows. There are hundreds of intermittent streams that flow briefly only during the frequent heavy rains.
Figure 1-7 Major catchments - Tutuila Island (study sites indicated) (from GIS Users Group, American Samoa Government, 2006)
Two-thirds of the land area of Tutuila has slopes > 17o (30%), and slopes greater than 45o are common along much of the coastline and in the uninhabited mountainous interior of the island. Consequently, the amount of cleared land for agriculture,
29 residences, and commerce is small compared to the total island land area. Habitation is almost exclusively confined to the southern coastal plain and southern coastal fringe.
Coral reefs are the dominant marine habitat for Tutuila near-shore waters, with 60% of the coastline occupied by narrow fringing reefs. Inhabitants of Tutuila routinely consume fish and shellfish from their village reefs, gleaning the reef flat and casting hook and line from the reef crest during periods of low tide, and spear-fishing on the fore reef mostly between dusk and dawn.
1.3.3 Climate
Climate for Tutuila is tropical, with consistent warm temperatures, high humidity, and abundant rainfall, throughout the year. Normal daily low and high temperatures are
24o and 30o C, with relative humidity ranging from a morning high of 86% to an afternoon low of 75%, each varying a few percent throughout the year (US National
Climatic Data Center, 2004). Annual rainfall on Tutuila is unevenly distributed due to the extreme topographic relief (Figure 1-8). At Pago Pago International Airport on the southeastern portion of the coastal plain, rainfall averages 3000 mm yr-1 (US National
Climatic Data Center, 2002). At Cape Matatula on the extreme eastern point of Tutuila mean annual rainfall is considerably less at 1800 mm yr-1 (Mefford, 2002). At higher elevations in the mountainous interior, average rainfall can exceed 5000 mm yr-1 (Izuka,
1999). Recent modeling studies (PRISM Group, 2006) suggest that some areas of the island may receive nearly 7 metres of rain annually, though recording stations have not been established in these areas for model validation (P. Peshut pers. obs., 2008).
Seasonal variations in rainfall are inconsistent due to the influence of the South
Pacific Convergence Zone (SPCZ) (Wright, 1963; Giambelluca et al., 1988). The
SPCZ is a regional phenomenon that is distinguished from the Inter-Tropical
30 Convergence Zone (ITCZ), the belt of circum-global low pressure that migrates roughly between the latitudes of 5o south and 15o north in the course of the solar year. On
Tutuila, the influence of the SPCZ results in increased cloud cover, elevated humidity, reduced variability in temperature, and increased rainfall, compared with tropical islands in similar latitudes and of similar topography (Giambelluca et al., 1988).
Figure 1-8 Rainfall distribution - Tutuila Island (study sites indicated) (from PRISM Group, 2006)
Precipitation on Tutuila occurs mainly in relatively short and intense rain squalls due to orographic processes, although prolonged periods of rain may occur during tropical cyclone (TC) events. Landfalls of tropical cyclones are relatively rare for the American
Samoa islands, with an average occurrence of < 1 per decade (US NOAA, 2005). The most recent TC landfall for Tutuila was TC Val in 1991 (Fiji Meteorological Service,
31 2007). Cyclone bypasses of a few hundred kilometres or more are more common than
TC landfalls, occurring on the order of once every few years (Fiji Meteorological
Service, 2007).
Prevailing winds on Tutuila are from the easterly quadrant throughout most of the year, and are generally weaker and more variable during the non-tradewind months of
December through April (Mefford, 2002; Mefford, 2004; US National Climatic Data
Center, 2004).
1.3.4 Geology and Petrography
In the published literature, the most complete and comprehensive descriptions for the formation, geologic base, and characteristics of the rocks of Tutuila are the works of
Daly (1924), MacDonald (1944) and Stearns (1944). Except as otherwise noted, these sources are summarised briefly below to give the reader a general overview of the
Tutuila volcanics.
Tutuila was formed from five principal volcanic centres that erupted in sequence during the Pliocene ( 5 mya) to earliest Pleistocene (2 mya). Successive broad shield volcanoes were generated from a tectonic plate hotspot located ~200 km to the east of
Tutuila’s present day position. The hotspost origin of the Samoa Islands was discovered in 1975 and was recently named the Vailulu’u Seamount (Staudigel et al.,
2006).
Volcanic centres on Tutuila occur over a distance of <30 km. Although the age relation of the rocks for each eruption centre has not been investigated, the sequence of eruption has been established, and each centre has been named for the convenience of discussion. The oldest eruption centres are the Masefau, Olomoana, and Alofau volcanos located in the eastern third of Tutuila. The younger Pago Volcano is centrally
32 located, and the Taputapu Volcano, the youngest, is westernmost. This general pattern of east-to-west formation for Tutuila is inconsistent with the westward movement of the
Pacific tectonic plate, and is likely a chance coincidence of close juxtaposition of eruption centres. The four oldest volcanos are highly eroded, and the youngest much less so, which may indicate that relative age differences among centres are not uniform.
Volcanos’ age, sequence of eruption, and weathered state, appear to be reflected in the present-day shape of the island (e.g., Figure 1-6), though this too may be coincidence.
The geologic base of Tutuila is chiefly volcanic basalts, with olivine basalts
(magnesium-iron silicates) of varying richness predominating. Volcanic ash, cinders and tuff also occur. Undifferentiated and unconsolidated sedimentary rocks of recent geologic age are found to depths of ~60 m in the valley floors and the talus slopes.
Calcareous sands and coral rubble mixed with small amounts of weathered basalts, and some cemented beach rocks, are found intermittently along coastal areas and are also of recent geologic age. Soils of Tutuila range from clayey to loamy, with varying amounts of silt; clay or loam characteristics are generally dependent on basalt or cinder parent material, respectively (USDA, 1983).
33 CHAPTER 2. RESEARCH OBJECTIVES and STUDY PLANS
2.1 Atmospheric Wet Deposition of Hg in the Remote Southern Hemisphere
Of principal importance in this work was to determine wet deposition of Hg on
Tutuila Island, as a means to examine how atmospheric deposition at a location in the remote southern hemisphere compares with suitably remote locations in the northern hemisphere. Quantification of Hg in Tutuila rainfall also provided a baseline for comparisons with future studies on Tutuila, or similarly remote sites, to assess changes in atmospheric Hg wet deposition at remote locations over time. In the future, long- term monitoring records from Tutuila or other remote southern hemisphere locations may be useful to describe changes in the global atmospheric reservoir of Hg, or the effectiveness of regulatory initiatives to control Hg emissions.
Mercury in Tutuila rainfall was also important to test model predictions for Hg in precipitation and Hg wet deposition rates for remote global locations. A prevailing argument against stricter regulatory controls on Hg uses and emissions is that contemporary experimental data for remote global locations is insufficient to reliably describe spatial or temporal trends for global atmospheric distribution and wet deposition of Hg (e.g., Jackson, 1997). In most efforts to substantiate concerns for the global relevance of Hg pollution, modeling currently prevails as the principal technique used to describe the occurrence of Hg in remote ecosystems. Modeling, without validation and support from experimental studies, lacks force as convincing evidence.
Arguments that the globalisation of Hg is of immediate concern are weakened without model validation by direct experimental measurements of Hg in precipitation in remote global locations. Comparisons of results from this research with modeling predictions
34 for Hg in rainfall and Hg wet deposition rates may help improve modeling predictability, and strengthen the use of models as a research tool.
Because of its small land area compared with the expanse of surrounding ocean, and the heavily forested terrain that limits locally generated aeolian dust that could un- accountably bias the measurement of Hg in precipitation, Tutuila effectively represented a stationary oceanic sampling site for Hg in remote southern hemisphere precipitation.
Specific research objectives for investigating Hg in Tutuila precipitation included:
Determine a reliable estimate of the mean and range of THg and MeHg
concentration in Tutuila rainfall, based on repeated measurements at a single
sampling station, for an extended-term sampling period, using rigorous ultra-
trace elements field and analytical techniques;
Examine differences in Hg concentration in Tutuila rainfall with contemporary
experimental data from comparably remote northern hemisphere locations, to
describe inter-hemispheric differences of Hg in precipitation at sites that are not
under significant influence of local or regional Hg emissions;
Determine annual wet deposition rates of Hg to selected Tutuila catchments on a
unit area basis and total mass basis;
Test predictions of current global models for Hg in rainfall and Hg wet
deposition rates for the Tutuila region of the South Pacific, using the
experimentally-derived data for Hg in Tutuila rainfall.
2.1.1 Study Plan - Atmospheric Wet Deposition
The sampling location in Alega Village was selected based on adequate rainfall events, topography, suitable clearances and windbreaks, and convenient access. An
35 underlying assumption for this work was that the atmospheric concentration of Hg is uniformly distributed over the expanse of ocean area surrounding the small land mass of
Tutuila Island, and therefore, mean concentration of Hg in rainfall on the windward shore of the Alega Village station was representative of Hg in rainfall for Tutuila overall.
In the oceanic region of Tutuila, limited experimental data indicates that atmospheric concentration of Hg is generally uniform at ~1 ng m-3 (Fitzgerald, 1995) which is in agreement with current global model predictions for the remote South Pacific
(Sunderland and Mason, 2007; Selin et al., 2008). At ultra-trace concentrations, error or variability in experimental results for rainfall samples might be introduced by inconsistency in maintaining ultra-clean field protocols, variability in analytical accuracy and precision, and the natural uneven distribution of rare particles in a turbulent medium.
Processes that lead to Hg wet deposition are known to be complex and are influenced by atmospheric conditions over the short and long term (Stein et al., 1996;
Lin and Pehkonen, 1999), which can introduce variability in experimental results. To account for seasonal variability, a sampling period of 12 consecutive months was selected, with samples collected twice per month, spaced approximately evenly at two week intervals. Side-by-side sampling apparatuses were deployed for one of the two monthly sampling events each month to evaluate precision in analytical methods and laboratory QA/QC procedures, and as a check on consistency for maintaining ultra- clean protocols for field activities.
Monitoring rainfall at the Alega station was intended to validate the rainfall distribution atlas that was recently completed for American Samoa (PRISM Group,
2006) to evaluate confidence in the overall atlas predictability for rainfall distribution
36 across Tutuila. Establishing confidence in the rainfall atlas was important because the atlas was used as the basis for calculating estimated Hg wet deposition rates for selected catchment areas and for Tutuila overall, based on the experimentally-derived mean concentration of Hg in Alega rainfall.
2.2 Distribution of Hg, Selected Elements, and Organic Carbon, in Marine Sediments, Stream Suspended Solids, and Upland Soils in the Tutuila Environment
Prior to undertaking this research there was no information on the occurrence or transport of Hg in the terrestrial or aquatic environments of Tutuila that could serve as a basis for the selection of study sites to examine the distribution of Hg among un- impacted and impacted environments, or to account for the potential of a natural source of Hg to aquatic systems.
Upland soils, suspended material in perennial streams, and marine sediments, were each investigated for Hg to describe some of the distribution and transport characteristics of Hg for a selection of coastal catchments and their associated fringing reefs.
Similar to Hg in tropical rainfall, the occurrence of Hg in marine sediments of tropical reefs is largely unknown for remote global locations (Wasserman et al., 2002).
Mercury in sediments from a selected geographic range of fringing reef sites was used as a “first-look” approximation of Hg distribution on Tutuila reefs. As a sink for pollutant inputs, and as a source for pollutant uptake by biota, sediments are widely recognized as an important component in the cycling of Hg in the aquatic marine environment. Physical characteristics, as well as the variable and complex chemical environments within the marine sediment matrix often serve to create conditions that are favorable for sequestration of a wide variety of pollutants (Kitamura et al., 1971;
37 Vonk and Sijpesteijn, 1973; Gianguzza et al., 2002). At the same time, abiotic and biotic processes within the upper sediment matrix, and at the sediment-water interface serve to cycle pollutant constituents, often in chemically altered forms, to the surrounding environment, and render them biologically available for uptake by organisms (e.g., Eckman, 1985; Meadows and Meadows, 1991; Aller, 1994; Gilbert et al., 1996; Gangaiya et al., 2001). In addition to Hg, selected common elements that were expected to occur in Tutuila marine sediments were investigated to evaluate sediment mineralogy (terrestrial vs. marine origin), to examine the potential for development of sediment quality guidelines based on the association of Hg with common terrestrial- or marine-derived elements.
On Tutuila, streams were assumed to be a significant transport mechanism of soil material between upland and aquatic environments. Erosion rates have not been evaluated for Tutuila, but estimates have been developed based on topography, soil types, and rainfall distribution (P. Peshut, un-published data, 2007). Estimates agree with observations that erosion rates on Tutuila are high, even in heavily forested pristine catchments where there is no land disturbance or human habitation (P. Peshut, pers. obs., 2001-2007). Mercury and total organic carbon (TOC) in stream suspended material were intended to provide a qualitative indication of terrestrial input of Hg to marine bays. Organic carbon was considered important because humic and fulvic acids of the soil humic material are known to bind Hg in soils, and humic material is thought to play an important role in transport of Hg via erosion processes (e.g., Semu et al.,
1986; Kainz et al., 2003; Xin and Gustin, 2007). Determining the relationship of TOC with Hg in stream suspended matter and reef sediments was intended to roughly characterise transport of Hg between the terrestrial and aquatic environments of Tutuila.
38 Upland soils were an important research focus because geochemical conditions of parent material can affect soil chemistry and composition (Ure and Berrow, 1982), and the geologic base could potentially be a significant contributing source of Hg in soils of remote environments. Geologic sources of Hg could confound the assessment of atmospheric wet deposition as a principal source of Hg to remote aquatic systems. It was therefore desirable to assess whether the geologic base was likely to be a significant factor in the accumulation of Hg in Tutuila soils. As discussed in Section 1.2.2, no data was available for Hg in Tutuila basalts. Soil Hg, compared to the probable range of Hg in oceanic basalts as described in the literature, was used to evaluate the potential of the geologic base as a significant source of Hg in Tutuila soils.
Specific research objectives to describe general distribution and transport aspects of
Hg for solid matrices on Tutuila included:
Compare soil Hg for Tutuila with concentrations of Hg in basalts as described in
the literature, to evaluate the potential for local geochemistry as a source of Hg
to Tutuila reefs;
Examine associations between Hg and TOC in stream suspended solids and in
reef sediments, and evaluate the distribution of Hg and TOC between these
compartments for each reef study site, to evaluate transport pathways for Hg
between upland and aquatic environments;
Document the mean and range for THg, MeHg, TOC, and selected elements in
stable sediments below the lower reef margin, among a geographic range of reef
study sites that represent a range of anthropogenic disturbance, to describe
distribution patterns of Hg in Tutuila reef sediments;
39 Based on observed distribution and transport patterns for Hg in soils, streams,
and reef sediments, select un-impacted and impacted study sites for research on
Hg accumulation in reef biota;
Evaluate the mineral composition of marine sediments to describe the
relationship between terrestrial- and marine-derived components in reef
sediments, to provide a qualitative assessment of the degree of terrestrial input
to Tutuila’s fringing reef systems, and to explore the potential for using common
elements in reef sediment as indicators of Hg contamination.
2.2.1 Study Plan - Sediments, Stream Suspended Solids, and Soils
For marine sediments, the objective was to sample stable undisturbed sediments as representative of long-term in situ conditions. Loose agglomerations that were subject to winnowing or re-suspension were not sampled; these criteria eliminated reef flat and fore reef sediments from the sampling plan (P. Peshut, pers. obs., 2001-2006). Tutuila reef flats and fore reefs are relatively high-energy environments, subject to sea conditions of exposed coasts. Observations showed that sediment distribution on
Tutuila reef flats was extremely patchy, the principal distribution characteristic being pockets of coarse coralline sand accumulated in reef moats. Fore reef sediments were observed to be more patchy than reef flat sediments, and occurred mainly as agglomerations of coarse particles in small pockets of the living reef matrix. Deep basin sediments at the terminus of the reef matrix were therefore selected to best represent long-term in situ conditions at each study site.
The number of samples from each study site was selected as the minimum number required to support adequately robust statistical analyses for comparisons of Hg occurrence among study sites. Sample size estimation was based on methods described
40 by Sokal and Rohlf (1981) and Zar (1999). This procedure proved only marginally useful. A sample size of n=32 was calculated, which indicated that the variance estimated from the known data was high.
Collection of reef sediments from all study sites over a short-term sampling period was considered important. The dynamic open coastal environment, and the relatively coarse-grained calcareous composition expected for Tutuila fringing reef sediments, was expected to result in stability and sequestration characteristics for contaminant constituents that were different than for sediments of lacustrine and estuarine systems.
Because sampling occurred over the short term, temporal variability in concentrations of Hg compounds within the sediment matrix was an unaccounted for unknown for this study component, although the potential was recognized.
Stream water samples were collected from the principal perennial stream for each of the selected study site catchments. Selection of sampling locations within each stream was based on eliminating any chance of seawater intrusion in stream water samples.
Sampling events were scheduled for periods of rainfall within each catchment area.
Water conditions on reef flats near stream outlets was used as a basis to roughly estimate consistency for stream sampling events in terms of rainfall. Rainfall duration and intensity were judged appropriate for stream water sampling when turbid water from the stream extended approximately 100 m into the clear waters of the reef flat.
This method was devised and judged appropriate as an estimation of comparability among sampling events because of the similarities of topography and stream hydraulic profiles for each study site catchment (P. Peshut, pers. obs., 2005).
Upland soils for this study were collected and analysed for THg as part of the
Military Munitions Response Program (MMRP), a USA government initiative under the direction of the Army Corps of Engineers (General Services Administration Schedule
41 No. GS-10F-0168J) and conducted by TLI Solutions, Inc., Golden Colorado, USA
(Delivery Order No. W91238-06-F-0042). For American Samoa, the MMRP initiative included the SFC Pele United States Armed Forces Reserve Center, which for programmatic purposes included the entire island of Tutuila and the small neighboring island of Aunu’u. The objective of the MMRP was to identify the presence of munitions constituents at U.S. military installations for the purpose of determining whether remediation actions were required. Mercury in upland soils was not part of
TLI’s original scope of work, because Hg was not an identified constituent in materials used for military training on Tutuila. Analyses for THg were included in the list of parameters at the author’s expressed request for the purposes of this study, under the auspices of the American Samoa Environmental Protection Agency.
Upland soil samples were collected from surficial soils in remote or rural locations only, to avoid potential anthropogenic influences from development or other economic activity. Land-use classifications were defined based on extensive field reconnaissance
(P. Peshut, pers. obs., 2001-2006). “Remote” areas were defined as areas without habitation, and no evidence of present or past land disturbance. “Rural” areas were essentially similar to remote areas, except for the presence of isolated residential structures.
2.3 Bio-accumulation of Hg in Tutuila Coral Reef Biota
Research objectives for this study component included investigations of Hg accumulation in reef plant and animal tissues, including three groups of fishes
(Acanthuridae, Mullidae, Sphyraenidae), and reef turf algae (species not determined).
The herbivorous accumulation pathway was represented by the Acanthurids
(surgeonfish), which feed exclusively on coral reef turf. The carnivorous accumulation
42 pathway was represented by the benthic-feeding Mullids (goatfish), and the top-trophic
Sphyraenids (barracuda).
For this study, it was assumed that turf algae and accumulated reef sediments are independent bases for feeding regimes among the culturally and economically important surgeonfish and goatfish. Turf algae are the exclusive food source for the lined surgeonfish Acanthurus lineatus, which was selected for the herbivorous pathway in this study (Craig, 1996; Myers, 1999; Randall, 2005). The goatfish (Mullidae) feed extensively on invertebrate benthic infauna (mainly worms, mollusks, and crustaceans) primarily in sediments below the lower reef margin (Wahbeh, 1992; Kulbicki et al.,
2005) and are potentially good indicators of uptake pathways from Hg sequestered in sediments.
Although the blackfin barracuda (Sphyraena qenie) is not important culturally or economically for Tutuila fisheries, this reef predator was selected to represent a top- trophic piscivore, as a proxy to reef predator species such as trevallys (Carangidae) and snappers (Lutjanidae), because of the relative ease of capture of barracuda compared to the latter fishes.
Fish groups represented varying degrees of site fidelity. Barracuda typically school on fore reefs during the daytime, and disperse to open water at dusk to feed (Myers,
1999; Randall, 2005). Compared with barracuda, the Mullids are not known to be far- ranging in foraging, and based on some evidence of seasonality of food supply
(Wahbeh, 1992) probably limit forays among sandy patches within relatively narrow ranges of less than a kilometre. In constrast to Sphyraenids and Mullids, A. lineatus are highly site specific (Craig, 1996) and defend feeding territories on the scale of square metres of turf algae, and are thought to move little beyond their home reef patches throughout their lifetimes.
43 Algal turfs of tropical coral reefs are distinguished from the dense mat-forming macroalgae turfs of temperate shores (Dahl, 1974; Morrissey, 1980; Hay, 1981,
Hackney and Sze, 1988). On coral reefs, multi-species micro-algal turfs are composed largely of uni-cellular and filamentous forms represented by all five major algal divisions, which dominate the reef plant community and provide up to 80% of total reef primary production (Adey and Steneck, 1985; Adey and Goertemiller, 1987; Klumpp and McKinnon, 1989; Williams and Carpenter, 1990; Van den Hoek et al., 1995;
Lobban and Harrison, 1997). Canopy heights in un-grazed areas are typically ≤ 4 mm, and in grazed areas ~ 1-2 mm (Wanders, 1977; Lobban and Harrison, 1997). As the major source of primary production on the reef, turf algae are grazed extensively by a great variety of marine herbivores at the micro-, meso- and macro-scales (Lobban and
Harrison, 1997). Grazing density is highest by micro-mollusks, with large numbers of individuals grazing per unit area of turf. At the meso-scale, grazing density is comparatively reduced from the micro-scale and is dominated by echinoderms (mainly urchins) and gastropods. Macro-scale grazing has the lowest density of individuals per unit area of turf, and is the exclusive province of fishes, and is dominated by the herbivorous members of the Pomacentridae (damselfishes), Chaetodontidae
(butterflyfishes) and Acanthuridae (surgeonfishes) (Myers, 1999; Randall, 2005). To compensate for intense grazing pressure, the adaptive mechanism for coral reef turfs is rapid growth to replace vegetative and reproductive tissue, with biomass turnover generally in the range of 4-12 days (Carpenter, 1985, 1986; Klumpp et al., 1987).
There was no data found in the published literature for accumulation of Hg in coral reef turf algae. Conversely, the occurrence of trace metals in macroalgae and associated biological processes involved for uptake are well documented because of extensive interest in macroalgae as bioindicators of trace metal pollution (e.g., Morris and Bale,
44 1975; Phillips, 1977; Denton et al., 2006). Microalgae assemblages that compose coral reef turfs have not been subjected to the same scrutiny.
Specific research objectives to evaluate bio-accumulation in Tutuila coral reef biota included:
Document the mean and range of THg and MeHg in marine water of selected
un-impacted and impacted reef sites, to serve as the abiotic reference datum for
calculating Bio-accumulation Factors (BAFs) for reef biota;
Document the concentration of THg and MeHg in turf algae, the exclusive food
source for A. lineatus, to improve resolution of trophic step bio-accumulation,
and to evaluate turf algae as a potential trophic uptake step for Hg in the reef
environment;
Document mean and range of THg and MeHg in muscle tissue for coral reef
fish species from three trophic levels on fringing reefs, to evaluate the extent of
bio-accumulation in remote un-impacted and impacted environments, and to
serve as a basis for human health risk assessment for exposure to Hg in remote
artisanal fisheries.
2.3.1 Study Plan - Bio-accumulation
Marine water sampling was scheduled for the trade-wind season of May through
October. Seasonality of rainfall is not sharply defined on Tutuila, and heavy rains can occur at any time. Selecting the “dry” trade-wind period for marine water sampling was intended to improve the probability of minimal terrestrial influence. As an additional measure, sampling was limited to periods for which no rainfall had occurred in the catchment area of the selected reef for at least 5 days prior to sampling. The selection of the 5-day criterion was based on observations of rainfall patterns for Tutuila (P.
45 Peshut, pers. obs., 2001-2006). Rainfall for all parts of Tutuila is plentiful, and the 5- day criterion was judged to offer the greatest likelihood of adequate sampling opportunities, combined with conditions of minimal terrestrial influence.
Twelve marine water samples from each targeted reef was selected as the minimum number of samples required to support statistical analyses for Hg among study sites.
This figure was based on methods described by Sokal and Rohlf (1981) and Zar (1999), for calculating required sample size (n). Historical Hg data for Tutuila reef waters that was useful for this study was limited to THg data from one “far-field” station, used for a
Territorial water quality monitoring programme, located seaward from the entrance to
Pago Pago Harbour (P. Peshut, un-published data, 2001-2006). The far-field site is geographically representative of Tutuila fringing reefs in relatively un-impacted areas, though there is undoubtedly some influence from the harbour on water quality in this area. A minimum sample size of n=8 was calculated. To improve robustness of statistical analyses, a sample size of n=12 for each study site was selected, based on available time and budgets for sampling and analyses. Since the target study sites were not previously investigated, the additional samples were added to improve the quality of the data set.
Six marine water samples from each targeted study site were selected as the number of samples required to support a qualitative assessment of total suspended solids (TSS) in reef waters. There is a relatively large amount of historical TSS data available for
Tutuila, which indicates that ambient conditions are generally ~1 mg L-1 TSS for un- impacted near-shore waters (CH2M Hill, 2007; DiDonato et al., 2007). The selected sample size of n=6 for TSS for each study site was based on the historical data combined with constraints on time and material budgets for field sampling and laboratory analyses.
46 Turf algae was collected and analysed as a composite “turf”. Identification of taxa was not attempted. Turf algae on tropical reefs typically contain species from all five major alga divisions, many species being unicellular and colonial, and a tropical reef algal turf may contain more than 100 species within a very small area (C. Lobban,
University of Guam, pers. comm., 2007).
Since there was no data available for Hg in turf algae on Tutuila reefs, estimation of sample size using the variance of historic data sets was not possible. A sample size of n=10 was selected for each study site, based on extensive observations of the consistency of turf composition and distribution on Tutuila reefs (P. Peshut, pers. obs.,
2001-2007), and professional judgment for adequate number of samples for statistical analyses, within constraints of time and material budgets.
Amongst the Samoa Islands, A. lineatus (alogo in Samoan) has an important economic and cultural role in the artisanal reef-fish fishery. Because of its cultural importance and favour as a food fish, A. lineatus represents an important exposure pathway for Hg in the human population. Surgeonfish are fecund and abundant, occupy relatively shallow-water habitats that are easily fished with a minimum of inexpensive gear, and are readily harvested when quiescent at night. The abundance and relative ease of capture of A. lineatus allowed for a balanced, single-species sampling design for herbivorous fish among study sites. A. lineatus are known to feed from ~10:00 a.m. until around dusk and then shelter until dawn (Craig, 1996). Capture was planned for the period 6:00-9:00 p.m. High site fidelity for surgeonfish improved the likelihood that the body burden of Hg in A. lineatus reflected ambient site conditions for Hg in the turf algae food source.
For the carnivores, the mid-level carnivore Mullidae (iasina in Samoan) are also a preferred food fish among the Samoa Islands, and are prominent in the local markets
47 and the artisanal catch. Abundance for Mullids is far less than for A. lineatus, so a single-species or single-genus sampling design was doubtful, though the plan was to collect as many of the same species as practicable within budget and time constraints.
Mullid feeding activity varies among species and genera, so both day-time and night- time collection forays were planned. Mullid collection was to be limited to areas of basin sediments, within the same general region of marine sediments collections.
Because of the limited foraging range for Mullids, their body burden of Hg was assumed to represent ambient site conditions for Hg in basin sediments as long as fish were collected from within the general limits of the bay site.
The top-trophic barracuda Sphyraena qenie (sapatū in Samoan) was selected as the target species to represent the top predators on Tutuila coral reefs. Expected abundance for S. qenie was the lowest of the fish selected for this study. They were also expected to be the most difficult to capture, being comparatively rare, and known to be taken only on hook and line; their speed and wariness preclude capture by net or spear. S. qenie are most often taken on a trolled lure, and are the most common Sphyraenid landed in American Samoa (P. Peshut, pers. obs., 2001-2007). Life histories for
Sphyraenids are not well known, but they are assumed to range widely among fringing reefs. Body burden for Hg in S. qenie was therefore not considered representative of a specific reef site on Tutuila, but was considered to be generally representative of top predator uptake of Hg in near-shore waters. The sole criteria for collection was that fish were captured in open coastal waters, at a suitable distance from Pago Pago Harbour, preferably the extreme southeastern, southwestern, or northern shores of Tutuila.
48 2.4 Human Health Risk Assessment for Hg in a Remote Artisanal Fishery
Concentrations of Hg in reef fish tissue as determined from the bio-accumulation study component (see Section 2.3) were intended to provide the basis for a human health risk assessment for surgeonfish and goatfish, two important groups of fishes in the remote southern hemisphere artisanal fisheries. Tutuila fringing reefs support a small but important commercial reef-fish fishery, and an extensive village-based fishery. Both fisheries rely exclusively on artisanal methods of capture, mainly net and spear, and to a much lesser extent, hook-and-line. Participation in these fisheries is traditionally of great socio-economic and cultural importance among Polynesians
(Meade, 1928). Even today, skilled artisanal fishers are esteemed throughout the
Pacific Islands, despite the spread of industrial-style economies and western cultural influences (P. Peshut, pers. obs., 1995-2008).
The purpose of the risk assessment component was to use experimental data for
MeHg in muscle tissue to evaluate health risks from exposure to Hg through consumption of Tutuila reef-fish, as an indication of Hg impacts in a remote artisanal fisheries where there are no identified sources of Hg.
Specific research objectives for the risk assessment were to:
Establish consumption rates for surgeonfish and goatfish from un-impacted and
impacted reefs in the remote global location of Tutuila, based on MeHg
concentration in muscle tissue, using established human health risk assessment
protocols;
Evaluate the assumption, as used in risk assessment protocols, that MeHg occurs
as 100% of THg in fish muscle tissue.
49 2.4.1 Study Plan - Risk Assessment
Risk assessment methodologies were planned in accordance with established US
EPA standard protocols and practice (US EPA, 2000). Concentration of MeHg in muscle tissue of selected reef-fish species from the bio-accumulation component of this study (see Section 2.3) was used as the basis to prepare consumption limit tables.
50 CHAPTER 3 . MATERIALS AND METHODS
3.1 Introduction
Methods employed for this study and materials and equipment used are organised in this chapter according to similarities of the environmental matrices investigated (water, solids, and tissue). Matrices within these categories share common analytical methods, common units for reporting, and generally similar field techniques for sample collection and handling.
All field activities, with the exception of upland soils (see Sections 2.2.1 and 3.5.3), were conducted by the author or under his direct supervision while in residence on
Tutuila Island, and while employed with the American Samoa Environmental
Protection Agency (2001-2008). Field work extended over a period of 27 months, with approximately 125 days of field activities, which included more than 100 dives using
SCUBA, to depths between 10-35 m. All field samples for rainwater (n=37), stream suspended solids (n=18), marine sediment (n=176), marine water (n=36), and turf algae
(n=30) were collected directly by the author with assistance from technical personnel in the field and at the on-island laboratory. All fishes (n=120) were collected by technical divers while under the author’s direct supervision in the field. All species identification and body measurements for fish were completed by the author with assistance from technical personnel. As discussed in Chapter 2, upland soil samples (n=62) were collected by an independent contractor for a separate and unrelated investigation, for which the author had direct involvement for study design, and for which he participated in approximately 30 field collections.
All sampling equipment was stringently acid-cleaned at the principal off-island analytical laboratory, and shipped and stored at the on-island staging laboratory in strict accordance with ultra-clean, trace elements field protocols (B. Lasorsa and G. Gill, pers.
51 comm., 2006, 2007). Collection, handling and shipping for all field samples were in accordance with the “clean hands, dirty hands” trace elements protocol, based on EPA
Method 1669 (US EPA, 1996b). Chain of custody was maintained for transport of all samples from the on-island laboratory to the analytical laboratories, and all samples were analysed within method specified holding times, unless otherwise noted.
All laboratory analyses for metals for all environmental matrices (except upland soils) were conducted by the Battelle Marine Sciences Laboratory, a part of the Pacific
Northwest National Laboratory, Sequim, Washington, USA. Total organic carbon in stream suspended solids and in marine sediments was analysed by Columbia Analytical
Services, Kelso, Washington, USA. Compositional analyses for mineral fractions for selected sub-samples of sediments were conducted at the School of Earth and
Environmental Sciences Laboratory, University of Wollongong, Australia. Total Hg in upland soil samples was analysed by Severn Trent Laboratories, Sacramento,
California, USA.
Mercury compounds in precipitation and marine water were analysed using cold vapour atomic fluorescence (CVAF), with analytical results reported as ng L-1. Metals in the solids matrices of upland soils, stream suspended solids, and marine sediments were analysed using cold vapour atomic absorbance (CVAA), CVAF, inductively coupled plasma mass spectrometry (ICP-MS), or inductively coupled plasma optical emissions spectroscopy (ICP-OES). All metals for solids matrices were reported as ng g-1 or μg g-1 dry-weight. For tissues, Hg compounds were analysed using CVAF and were reported as ng g-1 wet-weight.
Conventional wet chemistry procedures were used for analysis of non-metals. Total organic carbon for stream and marine sediments were reported as % dry-weight. Solids and lipids in tissue were reported as % wet-weight. Compositional analyses to
52 determine biogenous and terrigenous fractions of marine sediments used X-ray diffraction methods, with mineral composition reported as % dry-weight.
A summary table for analytes, instrumentation, analytical method, and achieved detection limit (DL) for each environmental matrix investigated is provided in relevant sections below. Summaries of analytical quality assurance and quality control (QA/QC) results are presented in Appendix A.
3.2 Statistical Applications for Evaluating Hg in Environmental Matrices in Accordance with Research Objectives
Statistical applications to evaluate analytical results were similar for each environmental matrix, throughout this work. In general, statistical applications included computation of descriptive statistics for each data set, and tests for homogeneity of variance and normality, to determine the applicability of parametric or non-parametric methods for selected comparisons among data sets. Distribution was tested for normality with the Shapiro-Wilk normality test. The Bartlett Test for homogeneity of variance was used when data was shown to be normally distributed. Hartley’s F-max
Test was used to test for homogeneity of variance if data were found to be non-normal, because the Bartlett Test is known to be sensitive to departures from normality (Zar,
1999). All field samples met criteria for independence and random sampling.
Analysis of Variance (ANOVA) was used for statistical comparisons if data met all underlying assumptions for parametric analyses. If ANOVA indicated a significant difference among data sets, the Student-Newman-Keuls multiple comparison of means test was used to determine where differences occurred. For comparisons among data that did not meet assumptions for ANOVA, the non-parametric Kruskal-Wallis
ANOVA was used (equivalently, Mann-Whitney U Test for comparisons limited to two data sets). When non-parametric tests indicated a significant difference among data sets
53 of equal sample size, the Student-Newman-Keuls means rank test was used to determine where differences occurred. For non-parametric analyses among data sets of unequal sample size, the Dunn procedure (Zar, 1999) was used to determine where differences occurred. All parametric and non-parametric statistical analyses were limited to single-factor comparisons.
Simple (two variable) linear regression was used to evaluate functional relationships between Hg and organic carbon for stream suspended solids and marine sediments, and between Hg in muscle tissue and body weight for fish. Significance of regression was evaluated by ANOVA.
Results for 15 selected elements from marine sediments sub-samples were pooled for statistical evaluation. Relationships among the concentrations of selected elements and
THg and MeHg were evaluated via a correlation matrix using the Spearman Rank
Correlation Coefficient (ρ). The Spearman method was selected over the Pearson
Product-Moment Correlation Coefficient as a non-parametric analysis, because the data set did not meet parametric assumptions for normality and homogeneity of variance.
The Spearman method had two important advantages over the Pearson method for this analysis. First, Spearman does not require that the relationship between variables is linear. Second, Spearman does not require the use of interval scale data, but can be used with ordinal scale data. Non-linearity between variables and ordinal scale data are characteristics of the data set for the selected elements in marine sediment sub-samples.
A significance level of α=0.05 was selected for all statistical analyses.
54 3.3 Study Sites and Sampling Stations; Un-impacted and Impacted Coral Reefs of Tutuila
Since there were no data available from which to discern patterns of Hg distribution among Tutuila coastal reef environments, the selection of advantageous and representative study sites to meet research objectives was at first largely intuitive (P.
Peshut, pers. obs., 2001-2006). Six marine bay sites (Figure 3-1) with similar fringing reef habitats were selected, of which, five sites represented fringing reefs on exposed coasts associated with catchments with minimal or no anthropogenic disturbance (“un- impacted”), and one site represented a fringing reef on a protected coast where there is extensive residential and commercial development in the catchment (“impacted”).
Throughout this text the colloquial term “catchment” is used in place of “watershed” to designate the contributory land area for a specified waterbody, consistent with the interchangeability of these terms found widely throughout the international literature.
Figure 3-1 Study sites - Tutuila Island
55 General reef geomorphology was similar for all study sites. Typically, there is a back reef of narrow to moderate width (< 150 m), with moat, and a prominent reef crest composed primarily of crustose coralline algae. Seaward from the crest is a steep-faced fore reef that terminates at depths of ~10-25 m at a shallow-grade talus sand slope. An exception to general reef geomorphology in this study is Tafeu Cove, which has a poorly developed reef flat structure that extends just a few metres seaward from the near-vertical mountain sides of the bay’s catchment, with no moat and only an intermittent reef crest.
All study sites represented healthy and robust tropical reef habitats with species-rich assemblages of scleractinian corals, benthic and cryptic fauna, and fish. A brief description of geographic and demographic data for each study site is presented below, along with specific sampling stations for each study component.
3.3.1 Masausi Bay; Masausi Catchment
Masausi Bay (Figure 3-2) is located on the north-shore of Tutuila and is the easternmost site selected for this study. Habitation within the Masausi catchment is limited to a single small village of < 200 residents (US Department of Commerce,
2001). The village is exclusively residential, with no commercial activity or structures.
Catchment area is 1.565 km2, with maximum elevation along the catchment divide ~300 m. Two small perennial streams discharge at the east side (Panota Stream) and west side (Vaipito Stream) of the bay. By sea, Masausi Bay is ~25 km from the impacted study site within Pago Pago Harbour.
56 Figure 3-2 Study site - Masausi (Masausi Bay)
3.3.2 Alega Bay; Alega Catchment
Alega (Figure 3-3) is located ~7 km from the impacted (Loa) study site within Pago
Pago Harbour. Habitation within this catchment is limited to a single village of < 50 residents (P. Peshut, pers. obs., 2007) with dwellings dispersed along the south-shore arterial roadway. Catchment area is 1.313 km2. Maximum elevation along the catchment divide is ~350 m. Alega Stream is the only perennial stream, and discharges about mid-way along the bay shoreline.
3.3.3 Loa; Pago Pago Catchment
The Loa study site (Figures 3-4A and 3-4B) is located within Pago Pago Harbour on the south side of Tutuila. Throughout this work, “Loa” is the nominal designation for this Harbour site, based on its location on the reef of Leloaloa Village, within the Pago
57 Pago Harbour catchment. As a Harbour location, Loa represents the only site for this study that is within protected waters and under significant anthropogenic influences.
Pago Pago catchment is steep-sided throughout, bordered almost entirely by
Tutuila’s prominent central ridge line. Buildable land within the catchment is generally limited to a narrow coastal fringe, but there is encroachment on steep hillsides which exacerbates soil erosion and leads to occasional land slippage. Highest point of land along the catchment divide is Mt. Matafao at 653 m. Five perennial streams discharge at approximately equally spaced intervals along the Harbour shoreline.
Figure 3-3 Study site - Alega (Alega Bay)
Population centres along the north-central and east shores of the Harbour include the villages of Leloaloa, Aua, and Onesosopo, with total population ~2500 (P. Peshut, unpublished data, 2007). Residential development dominates this section of the
58 Harbour shoreline, interspersed with a small amount of retail commerce. There is no centralised sewer system for these villages, and non-point discharges of inadequately treated domestic and animal wastewater are a chronic environmental concern (P.
Peshut, unpublished data, 2007). Land disturbance in the villages results in frequent discharges of sediments to reefs, though sedimentation rates have not been described.
Water quality concerns are exacerbated by Harbour configuration. Except for the extreme Outer Harbour and the Transition Zone at the seaward opening, Pago Pago
Harbour receives limited flushing by open ocean water, being principally a wind-driven mixing regime, with limited tidal mixing, so that water residence time increases progressively from the outer to inner Harbour regions (US ACOE, 1979; Costa et al.,
2007). Residence time for seawater at the Loa study site, with average wind and tide exchange, is estimated at ~15 days (US ACOE, 1979).
Figure 3-4A Study site - Loa (Pago Pago Harbour)
59 Figure 3-4B Study site - Loa (Pago Pago Harbour)
3.3.4 Amalau Bay; Vatia Catchment
Amalau Bay is a north-shore study site located by sea 37 km from the impacted site in the Harbour (Figures 3-5A and 3-5B). Throughout this work, “Vatia” is the nominal designation for the study site of Amalau Bay, which is located in a sub-catchment on the east side of the major Vatia catchment. Amalau Bay is separated from the main population centre of Vatia Village (~300 residents, P. Peshut, pers. obs., 2007) by a high sharp-crested ridge (~325 m). Human habitation at Amalau Bay is limited to 3 persons. There is no commercial activity within the confines of the Amalau Bay sub- catchment. The Vatia catchment area is 4.889 km2. Maximum elevation along the catchment divide is ~440 m. A single perennial stream (unnamed) with very low base flow discharges at the east side of Amalau Bay.
60 Figure 3-5A Study site - Vatia (Amalau Bay)
Figure 3-5B Study site - Vatia (Amalau Bay)
61 3.3.5 Tafeu Cove; Tafeu Catchment
Tafeu is a pristine, un-inhabited catchment within the boundaries of the National
Park of American Samoa, and shares a common boundary with the Vatia catchment on the east, and the Pago Pago catchment on the south. Throughout this work, “Tafeu” is used nominally to designate the study site of Tafeu Cove (Figure 3-6). Tafeu catchment
(5.116 km2) has no level land, being composed entirely of the mountainside along ~6 km of the island’s central ridge that plunges steeply to the sea between Maugaotula
Peak on the east and Fatifati Mountain on the west. Highest point of land along the catchment divide is ~490 m. Two perennial streams (unnamed) discharge to Tafeu
Cove, and numerous springs discharge from the steep mountain side along the entire length of the catchment. As described earlier, there is little or no reef flat development within Tafeu Cove. By sea, Tafeu Cove is ~37 km from the Loa site in the Harbour.
Figure 3-6 Study site - Tafeu (Tafeu Cove)
62 3.3.6 Fagafue Bay; Aasu Catchment
Aasu is the westernmost catchment selected for this study, and occupies part of the highly eroded northeastern slopes of the Taputapu Volcano, the youngest of the Tutuila eruption centres. Throughout this work, “Aasu” is used nominally to designate the study site of Fagafue Bay along the central coast of the catchment.
An accurate assessment of the level of anthropogenic disturbance in the Aasu catchment is difficult to make compared with the other sites selected for this study.
Population for the Aasu catchment is indeterminate, typically given as ~1150 based on census data (US Department of Commerce, 2001), but this is subject to speculation, and is not consistent with field observations. Virtually all of the population resides at the extreme southern boundary (most inland) of the catchment, some 2000 m from the shoreline (P. Peshut, pers. obs., 2007), and at an elevation of ~350 m. Population data for Aasu is potentially misleading, since the shore-side village is no longer extant, having been abandoned by the early 1940s due to extreme isolation, according to anecdotal evidence. The principal population centre is the Village of Aoloau, which straddles the island’s central ridge line at ~425 m elevation, and overlaps the catchments that are contiguous with Aasu. The Aasu catchment is heavily forested and land disturbance throughout most of the catchment is minimal. Land disturbance is limited to family agricultural plots and to a few dwellings in the extreme upland reaches.
Fagafue Bay (Figures 3-7A and 3-7B) is located in a sub-catchment of Aasu, and is isolated from the main body of the Aasu catchment by a prominent ridge, heavily forested slopes, and rocky outcrops. The highest point of land along the Fagafue sub- catchment divide is ~425 m. By sea, Fagafue Bay is ~45 km from the impacted Loa site in Pago Pago Harbour.
63 Figure 3-7A Study site - Aasu (Fagafue Bay)
Figure 3-7B Study site - Aasu (Fagafue Bay)
64 3.4 Mercury in Rainfall and Marine Water
3.4.1 Mercury in Rainfall and Monitoring Rainfall Events
3.4.1.1 Rainfall sample collection
Rainfall samples were manually collected from February 2007 to January 2008 at the
Alega station (Figure 3-3). For each sampling event, a new sampling apparatus was deployed, as supplied by the analytical laboratory. Each sampling apparatus was constructed of a 1 L Teflon® bottle, Teflon® collar, and a polycarbonate funnel, and were similar to apparatuses used in the Florida Atmospheric Mercury Study (Guentzel et al., 1995; Landing et al., 1995; Landing et al., 1998) except that collars were connected directly to the sample bottles (no interconnecting tubing used). Funnel configurations were square (9 cm) and circular (12 cm and 15 cm) with preference given to funnels with larger cross-sectional areas to limit the deployment period. For each sampling event, samplers were deployed in pairs or singly on an aluminium mast at a height of ~8 m above the ground surface. Prior to deployment, each sample bottle was covered with 4 layers of opaque white plastic bag to limit ultra-violet oxidation of
MeHg. Minimum target volume per sample was 200 mL, to satisfy laboratory QA/QC protocols.
Each sample bottle was pre-charged with 1.0 mL HCl prior to deployment (low-Hg
12M HCl, J.T. Baker Instra-Analyzed®). After retrieval, each sample was preserved at
0.5% HCl, with additional acid required for preservation calculated based on the retrieved sample volume, less the pre-charged acid volume. Immediately after acid preservation, samples were stored in complete darkness in clean polypropylene coolers at cool ambient temperature until shipment to the analytical laboratory. Preserved samples were generally held less than one month before shipment to the analytical laboratory.
65 3.4.1.2 Analytes, analytical methods, and detection limits for Hg in rainfall
Rainfall samples were analysed for THg and MeHg as indicated in Table 3-1.
Analytical methods are performance based, and were modified to meet analytical laboratory DL and QA/QC standards, as summarised below. All reagents used in the analyses were verified to have very low Hg content or were purified prior to use to ensure negligible reagent blank concentrations.
Table 3-1 Methods and detection limits for Hg in rainfall Analyte Instrument Method1 Achieved DL2
Total Hg CVAF EPA 1631E 0.188 ng L-1
Methyl Hg CVAF EPA 1630 0.0159 - 0.0188 ng L-1
1(US EPA, 2001a, 2002) 2(Battelle Marine Sciences Laboratory, 2007, 2008)
For THg, all Hg in a 100 mL aliquot of pre-acidified field sample was oxidised to
Hg2+ by the addition of 0.5 mL BrCl solution. The BrCl oxidising solution was prepared from 27g KBr added to 2.5 L concentrated HCl, stirred for 1 h, after which 38 g KBrO3 was added and stirred for an additional 1 h. Following oxidation, the sample was sequentially reduced with 0.25 mL NH2OH•HCl to destroy free halogens.
Following destruction of free halogens, Hg2+ in the sample was reduced to Hg0 with 0.5 mL of SnCl2 solution (prepared from 200 g SnCl2•2 H20 and 100 mL concentrated HCl, brought to a final volume of 1.0 L with de-ionised water, then purged with Hg-free N2 gas overnight at 500 mL min-1). Elemental Hg was then purged onto a gold trap with
-1 N2 gas flowing at 350±50 mL min . Elemental Hg was thermally desorbed from the gold trap, trapped and subsequently desorbed from the analytical trap, then carried into the detector by Ar gas flowing at 120 mL min-1. Elemental Hg in the sample was quantified as THg by a CVAF detector (Tekran model 2500).
66 For MeHg, samples were distilled into clean water using “sending” and “receiving” vials. Each receiving vial was prepared with 5 mL de-ionised water, fitted with distillation cap, and immersed in an ice bath. Each sending vial was prepared with 50 mL of pre-acidified field sample, fitted with a distillation cap, and connected to the receiving vial with a 3 mm Teflon® tube. Pre-acidified samples required no additional reagents. The sending vial was placed in a heating block at 125o C and immediately
-1 connected to N2 gas flowing at 40 mL min . When the volume in the receiving vial reached 45 mL the distillate sample was disconnected and capped and stored in the dark until analysis (not more than 24 hours). For analysis, the distillate sample was placed in the cold vapour generator with 20-30 mL de-ionised water and 200-300 μL of acetate buffer and agitated to mix thoroughly. While mixing, 125 μL of 1% sodium tetraethylborate was added directly to the sample to form volatile methylethyl Hg. The sample was allowed to react for 17 minutes and then purged onto graphite carbon traps with Ar gas flowing at 200 mL min-1 to pre-concentrate and remove interferences.
Samples were then isothermally chromatographed, pyrolytically decomposed to Hg0, and quantified by the CVAF detector.
3.4.1.3 Blank correction for Hg in rainfall
One acid blank and two field blanks were prepared during the 12-month sampling period to account for potential sample contamination by Hg from acid used for preservation, from Hg-contaminated sampling equipment, or contamination from handling procedures during equipment deployment and retrieval. Acid and field blanks were prepared separately and were applied separately for blank correction of analytical results. The acid blank consisted of a ~25 mL aliquot of HCl taken from the 500 mL original acid container. Field blanks were prepared during the first quarter (March
67 2007) and last quarter (October 2007) of the rainfall sampling period. Field blanks were collected using a clean, newly deployed sampling apparatus. High purity DI water used for each field blank was supplied in a 1 L Teflon® bottle from the analytical laboratory. Each field blank was prepared by carefully and slowly pouring one-half of the DI water through the collection funnel to simulate a rainfall event. All interior surfaces of the funnel were wetted with the DI water. Field blank water and water from the sampler were then preserved with 0.5% HCl similar to rainfall samples. The field blank correction was calculated by taking the difference between the Hg concentration of the water which passed through the funnel and collar apparatus, and the Hg concentration of the original water which was not passed through the funnel. For field blanks, Hg in the acid used for preservation was accounted for. The mean of the two field blanks was used as the correction factor. Acid and field blank corrections were subtracted from analytical results prior to application of statistical analyses or any other data evaluation.
3.4.1.4 Monitoring rainfall events
Precipitation was monitored continuously at a single gauging station in proximity to the rainfall sampling station (Figure 3-3), from January 2007 through June 2008, using manual and automatic rain gauges. The manual gauge served principally as a means to estimate volume in rainfall sample bottles, based on depth of rainfall and funnel geometry. The manual gauge also served as a backup for the automatic gauge. Manual and automatic gauge readings were compared periodically during the study period to assess agreement of rainfall measurements. Rain gauge installations were discussed with US Geological Survey personnel who had completed similar rainfall measurement
68 studies for Tutuila (S.K. Izuka, pers. comm., 2006.), and were reviewed by the analytical laboratory (G. A. Gill, pers. comm., 2006).
Predictability of the rainfall distribution atlas for American Samoa (PRISM Group,
2006) for the Alega gauging station was evaluated based on 18 months of rainfall data, using 7 iterations of total rainfall for successive 12-month periods. Because of expected long-period variability in climatic patterns, it was not known whether the sampling period for this study occurred during a relatively wetter or drier period. This iterative method was used as an indicative measure of potential bias in the relatively short-term rainfall data collection period. Total rainfall for each 12-month period was compared to the atlas prediction for the Alega gauging station. Variability was evaluated as % deviation of experimental values from the atlas value.
3.4.1.5 Calculation of Hg wet deposition rates
Estimated annual wet deposition rates for each study site catchment area and Tutuila overall were calculated with assistance from Raymond Laine of the School of Earth and
Environmental Sciences, University of Wollongong, using ArcGIS 9.2® software.
Rates were calculated from the experimentally-derived concentration for Hg in Alega station rainfall, and mean annual rainfall volume for catchment areas. Rainfall volume was calculated using annual rainfall distribution polygons within each catchment, taken from the rainfall atlas digital files (PRISM Group, 2006), and land area beneath prescribed polygons. Land area data was obtained from the American Samoa GIS database (GIS Users Group, American Samoa Government, 2006).
69 3.4.2 Mercury in Marine Water
3.4.2.1 Marine water sample collection
Marine water samples were collected at Loa, Vatia and Aasu on three separate occasions during July, August, and October 2007. Sampling locations are shown in
Figures 3-4A, 3-5A and 3-7A. Two samples for Hg analyses and one sample for TSS analysis were collected from each of two sampling stations within each bay for each sampling event. Each sample was collected in a clean 1 L Teflon® bottle supplied by the analytical laboratory. Samples were collected by hand using SCUBA at a depth of
~8 m, and ~25 m seaward from the fore reef. For sample collection at depth, a capped bottle was taken to ~2 m, uncapped and flooded, then capped and taken to ~8 m. At the final sampling depth, the bottle was uncapped and flexed 10 times to ensure at least one complete exchange of water volume. After collection, sample bottles were immediately delivered to the cooler on the boat and stored on ice in complete darkness until return to the on-island laboratory.
At the on-island laboratory, marine water samples for Hg analyses were preserved
® with 0.2% low-Hg 18M H2SO4 (J.T. Baker Instra-Analyzed ). Immediately after preservation, samples were stored in complete darkness in clean polypropylene coolers at cool ambient temperature until shipment to the analytical laboratory. Samples for Hg analyses were generally held for one week before shipment to the analytical laboratory.
Marine water samples for TSS analysis were not acid preserved because of potential breakdown of filter membranes by the acid during analysis, which could lead to a bias in results. Samples for TSS were refrigerated at 4o C in total darkness until shipment to the analytical laboratory.
Holding time for TSS was exceeded by 3 days for the first sampling event, due to a delay in commercial flights from Pago Pago to Honolulu. Since samples were held in
70 complete darkness at 4o C during the entire storage and shipping period, and given the extremely low TSS in samples, it is not expected that results were significantly affected by this marginal exceedance.
3.4.2.2 Analytes, analytical methods, and detection limits for Hg in marine water
Analytical methods and laboratory procedures for THg and MeHg in marine water were identical to those for Hg in rainfall (see Section 3.4.1.2 and Table 3-2). For TSS,
1.0 L of un-preserved marine water sample was vacuum filtered on pre-weighed 1.0 μm filter membranes. Membranes were dried to constant weight at 60o C. Weight differential was calculated to obtain the final reported result.
Table 3-2 Methods and detection limits for Hg in marine water Analyte Instrument Method1 Achieved DL2
Total Hg CVAF EPA 1631E 0.188 ng L-1
Methyl Hg CVAF EPA 1630 0.0188 ng L-1
TSS SM2540D 0.1 mg L-1
1(US EPA, 2001a, 2002; Rice et al., 2006) 2(Battelle Marine Sciences Laboratory, 2007, 2008)
3.4.2.3 Blank correction for Hg in marine water
One acid blank and one field blank was prepared to account for potential contamination from Hg in acid used for marine water sample preservation, from contaminated sample bottles, or from contamination from handling in the field. The acid blank consisted of a ~25 mL aliquot of H2SO4 taken from the 500 mL original acid container. The field blank consisted of a sample bottle filled in the field from DI water provided by the analytical laboratory. Blanks were prepared, handled, and shipped in accordance with procedures and protocols described for the rainfall blanks (see Section
71 3.4.1.3). Correction factors for THg and MeHg were not required for marine water because the amount of Hg detected in the acid and field blanks was insignificant based on the volume of acid used.
3.5 Mercury in Marine Sediments, Stream Suspended Solids, and Upland Soils
3.5.1 Marine Sediments
3.5.1.1 Marine sediments sample collection
Marine sediment samples were collected by hand via SCUBA from stable bottom sediments, generally 1-2 m seaward from where the reef matrix terminated in sand at a shallow-grade talus slope. Samples were collected in certified clean, pre-labeled wide- mouthed trace metals 250 mL glass jars with Teflon®-lined lids, supplied by the analytical laboratory. Sample jars remained sealed and were stored in clean polypropylene coolers for transport to sampling locations (Figures 3-2 through 3-7A).
At the surface and well away from the boat, jars were opened, flooded, and re-sealed at a depth of ~1 m. At depth, sampling stations were spaced at approximately 10 m intervals. At each sampling station, a sample jar was opened and waved vigorously in the water column to clear surface water from the jar. Surficial sediment was collected by carefully dredging sediments with a jar held horizontal, being careful to dredge to a depth not greater than the width of the jar mouth (~6 cm). Dredging was done slowly and carefully so to not disturb sediments, and so that not more that ~¾ of the jar volume was filled. Jars were immediately sealed after sample collection. At the surface, samples were placed on ice in complete darkness for transport to the on-island laboratory.
At the on-island laboratory, jars were allowed to settle for ~2 h in complete darkness at 4o C. After settling, jars were carefully opened and water decanted by gently rolling 72 the jar at a shallow angle so that water slowly poured off at a steady rate over the entire circumference of the jar rim. This procedure served to remove excess water from the jar, and removed any loose sediment particles from the jar rim to ensure a tight seal.
Care was taken so that there was no loss of fines during decanting. After decanting, sediment samples were frozen at -20o C until shipment to the analytical laboratory.
3.5.1.2 Additional elements in marine sediments
In addition to THg and MeHg, 15 elements were analysed in 6 randomly selected sub-samples (random number generation techniques) taken from sediment samples from each study site. Sub-samples were taken from homogenised sediment at the analytical laboratory. Additional elements were evaluated to examine correlations between Hg compounds and commonly found elements in reef sediments (biogenous and terrigenous) to explore the potential for developing qualitative assessment techniques for Hg contamination based on relatively inexpensive analyses of commonly found elements. Additional (common) elements were selected based on expected occurrence in fringing reef sediments of tropical Pacific high islands (R.J. Morrison, pers. comm.,
2006).
3.5.1.3 Analytes, analytical methods, and detection limits for Hg and selected elements in marine sediments
Marine sediment samples were received frozen at the analytical laboratory and were stored at -80o C until processing. MeHg was analysed on wet sediment and the concentration was converted to a dry-weight basis using % moisture data. Following
MeHg analysis, samples were freeze-dried and homogenised to a fine powder in a ball mill prior to digestion for analyses of THg and selected elements, as indicated in Table
3-3.
73 Table 3-3 Methods and detection limits for Hg and elements in marine sediments Analyte Instrument Method1 Achieved DL2, dry-weight
Total Hg CVAF EPA 1631E 0.302 - 0.320 ng g-1
Methyl Hg CVAF EPA 1630 0.0124 - 0.0164 ng g-1
TOC ASTM D4129-82 0.05 %
Al ICP-OES EPA 200.7 2.03 - 2.28 μg g-1
As ICP-MS EPA 200.8 0.02 - 0.404 μg g-1
Ca ICP-OES EPA 200.7 0.850 - 1.24 μg g-1
Cd ICP-MS EPA 200.8 0.0040 - 0.0060 μg g-1
Cu ICP-OES EPA 200.7 0.0352 - 0.132 μg g-1
Fe ICP-OES EPA 200.7 0.385 - 0.944 μg g-1
K ICP-OES EPA 200.7 13.8 - 21.5 μg g-1
Mg ICP-OES EPA 200.7 1.66 - 2.12 μg g-1
Mn ICP-OES EPA 200.7 0.0286 - 0.047 μg g-1
Ni ICP-OES EPA 200.7 0.0434 - 0.0535 μg g-1
P ICP-OES EPA 200.7 2.72 μg g-1
Si ICP-OES EPA 200.7 16.5 μg g-1
Sr ICP-OES EPA 200.7 0.0459 - 0.0928 μg g-1
Ti ICP-OES EPA 200.7 0.0767 - 0.100 μg g-1
Zn ICP-OES EPA 200.7 0.0614 - 0.151 μg g-1 1(ASTM, 1982; US EPA, 1994a, 1994b, 2001a, 2001b, 2002) 2(Battelle Marine Sciences Laboratory, 2008; Columbia Analytical Services, 2008)
The atomic fluorescence techniques used for Hg analyses in marine sediments (EPA
1631E, EPA 1630) were similar to methods used for Hg analyses for rainfall and marine water samples, and differed only in the digestion of the solids matrix for analysis.
Performance-based method modifications are summarised below.
For sample digestion for THg analysis, 5 mL of aqua regia (4 HCl: 1 HNO3) was added to a 0.2 g aliquot of homogenised sample in a glass vial and allowed to digest at room temperature for 24 h. After digestion, 16 mL of a solution of 0.07N BrCl in DI water was added to destroy any remaining organics. Following digestion, Hg2+ in an
0 aliquot of the sample was reduced to Hg by the addition of SnCl2 solution to the purge vessel as described for the rainwater samples (see Section 3.4.1.2).
74 For MeHg, sample digestion was in accordance with procedures described for solids by Bloom et al. (1997). A 1.0 g aliquot of homogenised sample was placed in a Teflon® centrifuge tube with 1 mL of 1 M CuSO4 solution, 5 mL of acidic KBr solution, and 10 mL of methylene chloride. The samples were allowed to react for 1 h and then shaken for 1 h. Following shaking, the samples were centrifuged at 2000 rpm for 30 min.
After centrifugation, 2 mL of methylene chloride was pipetted from the extract and placed in 45 mL of DI water. The DI water was then heated to 125° C to boil away the methylene chloride, leaving the MeHg in the DI water. The water was then analysed as described for the rainwater samples (see Section 3.4.1.2).
Two digestions were used for ICP-MS and ICP-OES metals analyses.
Digestion 1 - Total Dissolution with Boric Acid Neutralisation: Approximately 300
® mg (dry-weight) of each sample was combined with HNO3, HCl, and HF in a Teflon bomb and heated in an oven at 130º C (±10º C) for a minimum of 8 h. After cooling, boric acid was added to the digestate to eliminate analytical interferences caused by HF interactions with select analytes, and DI water was added to achieve analysis volume.
Digested samples were analysed for Al, Ca, Fe, K, Mg, Mn, P, Si, Sr, and Ti.
Digestion 2 - Aqua Regia Digestion: Approximately 200 mg (dry-weight) aliquot of
® each sample was combined with HNO3 and HCl (aqua regia) in a Teflon bomb and heated in an oven at 130º C (±10º C) for a minimum of 8 h. After cooling, DI water was added to the sediment digestate to achieve analysis volume. Digestates were analysed for Cu, Ni, Zn, As, and Cd.
Digested samples were analysed according to Battelle SOP MSL-I-033
Determination of Elements in Aqueous and Digestate Samples by ICP-OES based on method EPA 6010B (US EPA, 1996c) and method EPA 200.7 (US EPA, 1994a) as modified and adapted for analysis of low-level samples, and according to Battelle SOP
75 MSL-I-022 Determination of Elements in Aqueous and Digestate Samples by ICP-MS based on method EPA 1638 (US EPA, 1996a) and method EPA 200.8 (US EPA, 1994b) as modified and adapted for analysis of low level sediment and tissue samples.
Samples were prepared for TOC analysis by the addition of HCl to remove inorganic carbon, dried to constant weight at 70º C, then homogenised to a fine powder in a ball mill. Samples were combusted in a quartz boat at ~1000º C in a pure oxygen stream at
~125 mL min-1. Combustion product gases were swept through a barium chromate catalyst/scrubber to ensure that all C was oxidised to CO2, then swept to the coulometer where C was quantified by automatic coulometric titration.
3.5.1.4 Mineralogical analysis for marine sediments
Mineral composition of reef sediments was evaluated as a qualitative measure of terrestrial material input to Tutuila reef systems. Mineral compositions of 6 sediment sub-samples randomly selected from each study site (see Section 3.5.1.2) were analysed by X-ray diffraction to determine the proportions of calcite and aragonite in the carbonate fraction in marine sediments (Morrison et al., 2001). Additional minerals evaluated included quartz, kaolinite, illite, chlorite and halite. The proportions of the carbonates to the non-carbonate minerals provided an estimate of the biogenous and terrigenous fractions in Tutuila reef sediments at selected study sites.
3.5.1.5 Estimation of reduction/oxidation (redox) horizon in marine sediments
Redox horizon was evaluated for reef sediments to show that sediment samples were obtained consistently from the oxic sediment layer among reef study sites. Two methods were attempted to estimate the depth of the redox horizon. First, sediment samples were brought to the surface in closed containers and analysed with a hand-held
76 meter with platinum probe. This proved impractical, because exposure of sediments to oxygen in the water column after disturbance from in situ conditions, and then exposure to oxygen in the air during meter readings, resulted in rapidly changing redox conditions, and no stable or reliable reading could be obtained. The second method employed was a qualitative approach. The redox horizon for in situ sediments was estimated by identifying the colour change that is often associated with the change in conditions from a reducing to an oxidizing environment. It was assumed that sediments of the Tutuila fringing reef environment followed typical colouration patterns, where sediments in the oxidised strata are whitish to light brown, and sediments in the reducing environment of anoxic sediments are dark brown, varying shades of gray, or nearly black.
Stations were selected along the same transect established for marine sediment sampling. Colour change within the sediment strata was assessed using a clear PVC core tube (50 cm x 5 cm) fitted with soft plastic endcaps. The uncapped open core barrel was worked slowly into the sediment until refusal. At refusal, the upper end of the barrel was capped. The core barrel was then worked slowly out of the sediment and carefully placed on the adjacent solid sand surface to avoid disturbance of the core.
For each core a distinct line of colour change was visible, although the contrast and definition varied among cores. The depth of colour change in the core was measured directly to 0.1 cm with a fine-line stainless steel rule. In most cores, the colour change zone was sharp and distinct (≤ 0.5 cm in width). If the colour change appeared gradual, the centre of the colour change zone was assumed to represent the redox horizon.
77 3.5.2 Mercury in Stream Suspended Solids
3.5.2.1 Stream water suspended solids sample collection
Three stream water samples were collected from the principal perennial stream for each of the study sites between July 2006 and January 2008 (Figures 3-2 through 3-7A).
Stream water was collected in clean PVC containers with drip-tight lids. Containers were purchased new for the purpose of this study and were not used for any other purpose throughout the sampling period. All containers were of identical size (5 gallon, nominal) and from the same manufacturer. Prior to initial and subsequent deployment to the field for sampling, each container and lid was detergent washed 3 times and allowed to air dry in an air-conditioned room. After drying, lids were put tightly in place and were not removed until the time of sampling in the field. Containers and associated lids were appropriately labeled and dedicated for each study site, and no cross-over between containers or lids, and containers and sites, occurred.
To ensure there was no intrusion of seawater, stream water samples were collected within 50 m upstream of the stream outlet, at or near the time of low tide, which coincided with an elevation of 1-3 m above the obvious high water elevation for the bay. Samples were collected mid-stream in a depth of water that allowed for complete submersion of the container in the stream, with suitable clearance from the stream bottom. Containers were positioned upstream, with the opening in the direction of on- coming current. The lid was put tightly in place while container remained submerged, so that no leakage occurred. Lids remained intact for transportation to the on-island laboratory and for the duration of the settling period.
At the on-island laboratory, sealed containers were left undisturbed for 28 days to facilitate settling. Following the settling period, lids were carefully removed and water was decanted via siphoning using clean 6-mm PVC tubing. Following decanting,
78 sediment was washed from sides and bottom of each container using DI or distilled water to consolidate sediment for transfer to trace metals clean 250 mL glass jar with
Teflon®-lined lids, supplied by the analytical laboratory. The sediment slurry was frozen at -20o C until shipment. Upon arrival at the analytical laboratory, the samples were again allowed to settle and overlying water was decanted and pipetted off until only the solids and a small amount of water remained. Samples were then freeze-dried.
3.5.2.2 Analytes, analytical methods, and detection limits for Hg and TOC in stream suspended solids
Analytical methods and procedures for TOC in stream suspended solids were identical to those used for TOC in marine sediments (see Section 3.5.1.3). For THg, samples were prepared for analysis as discussed for marine sediments (see Section
3.5.1.3) and then analysed by cold vapour atomic absorption techniques by reacting an
0 aliquot of the digestate with SnCl2 solution to reduce all Hg in the sample to Hg , then sweeping it out of the sample through the CVAA detector (Table 3-4).
Table 3-4 Methods and detection limits for Hg and TOC in stream suspended solids Analyte Instrument Method1 Achieved DL2, dry-weight
Total Hg CVAA EPA 245.5 3.00 ng g-1
TOC ASTM D4129-82 0.05%
1(ASTM, 1982; US EPA, 2007) 2(Battelle Marine Sciences Laboratory, 2008; Columbia Analytical Services, 2008)
3.5.2.3 Blank correction for Hg in stream suspended solids
De-ionised or distilled water that was used to rinse and consolidate settled stream solids from container sides and bottom was analysed for THg prior to use to determine potential contamination of samples with THg in rinse water. Four DI or distilled water
79 sources were available for use at the on-island laboratory, labeled DB1-DB4 for THg blank analyses. De-ionised water was obtained from on-island laboratory stock
(Protocol®, DB1). De-ionised water was also obtained from the quality control laboratory of a local food processing industry (Barnstead MegaPure System®, DB2).
Commercially available distilled water (Arrowhead®, DB3) was purchased at a local retailer, and trace metals clean blank water was obtained from Battelle Marine Sciences
Laboratory (DB4). A 500 mL sample of each water source was placed in a clean 1 L
Teflon® bottle supplied by the analytical laboratory, and was handled, shipped, and analysed according to procedures described for rainfall and marine water samples. De- ionised water source, and volume used for processing each stream water sample are shown in the results (Chapter 4).
For a conservative THg blank correction for stream suspended solids, it was assumed that all Hg present in the volume of rinse water that was used for processing each stream sample was adsorbed onto the sediment in the slurry. The mass of THg added to the sediment slurry was calculated from the Hg concentration in the water and the volume of water used, then subtracted from the concentration reported by the analytical laboratory.
3.5.3 Mercury in Upland Soils
3.5.3.1 Upland soils sample collection
Sampling stations for upland soils included dry drainage ways and non-drainage way areas. Numerous dry drainage ways occur on Tutuila uplands due to extreme topographic relief, and are subject to intermittent flows during rains. To distinguish sampling locations, soil collected from dry drainages was labeled “SED”, and soil collected from non-drainage areas was labeled “SO” (see Table 4-21). All sampling
80 locations were geo-located with hand-held GPS units that were calibrated with a central station established on Tutuila (Figure 3-8).
Figure 3-8. Upland soils sampling stations
For non-drainage way areas, soil was collected in a radial perimeter pattern (“spoke and hub”) and composited. Five discrete samples were collected from points within a central area of approximately 30 cm square, and 6 discrete samples were collected from approximately evenly spaced points along the circumference of a circle with radius approximately 30 cm from the edge of the central area. Each discrete sample was taken from a depth of 10-15 cm below vegetation cover (if present) using a clean disposable plastic trowel. Soil samples were deposited and sealed in a clean zip-seal bag and composited by kneading until the sample was completely homogenised. Homogenised samples were transferred to 250 mL trace-metals clean glass jars with Teflon®-lined
81 lids, and stored on ice in complete darkness for transport to the analytical laboratory.
Preservation of soil samples for THg analysis was not required, in accordance with the analytical method (US EPA, 1994c).
For sampling locations within dry drainage ways, discrete samples were collected in a linear pattern upstream and downstream along the drainage way center-line.
Sampling equipment and field protocols were identical to those used for non-drainage locations, except that discrete samples were taken at 3 locations, spaced linearly ~1 m apart.
3.5.3.2 Analytes, analytical methods, and detection limits for Hg in upland soils
Method 7471A (Table 3-5) is an atomic absorption technique that differs from THg analysis in rainfall and marine water samples only in the digestion of the solids matrix and the instrumentation used (absorbance vs. fluorescence).
Table 3-5 Methods and detection limits for Hg in upland soils Analyte Instrument Method1 Achieved DL2, dry-weight
Total Hg CVAA EPA 7471A 0.0092 - 0.0210 mg kg-1
1(US EPA, 1994c) 2(Severn Trent Laboratories, 2007)
For sample digestion, 5 mL reagent water and 5 mL aqua regia (4 HCl: 1 HNO3) were added to 0.2 g aliquot of homogenised sample in a BOD bottle and heated for 2 min in a 95o C water bath. After cooling, the sample was digested by the addition of 50 mL of reagent water and 15 mL of KMnO4 solution, mixed thoroughly, and heated for
o 30 min in a water bath at 95 C. After cooling, 6 mL NH2OH•HCl was added to destroy free halogens. Following destruction of free halogens, Hg2+ in the sample was reduced
0 to Hg by the addition of 5 mL of SnCl2 solution and 50 mL of reagent water.
82 Elemental Hg was aerated from solution in a closed system by Ar gas flowing at 1 L min-1, and quantified as THg by the CVAA detector.
3.6 Mercury in Coral Reef Biota
3.6.1 Turf Algae Collection
Collection stations for turf algae were spaced at ~30 m intervals (Figures 3-4B, 3-5B and 3-7B) based on a selected transect length that approximated the transect length for marine sediment collections, and which encompassed the length of reef where herbivorous fish were collected. All turf algae patches sampled were qualitatively selected for similarity in color, texture, density of cover, and growth height. Samples were collected in a depth range of 5-10 m at all locations, which encompassed the grazing range of the herbivorous fish.
Un-grazed turf (relatively) was selected over turf from grazed areas. The extent of grazing pressure was determined on the basis of clear and apparent excessive growth, compared to the obviously grazed stands. Un-grazed turf was selected over grazed turf to facilitate collection of adequate tissue mass for ultra-trace elements analyses, since grazed turf was too closely cropped for efficient collection.
Turf algae were collected by dislodging turf-covered reef substrate by hand or with a small stainless steel hammer. Small- to medium-sized substrate projections of ~15-20 cm in length, with a length-width ratio of approximately 4:1, were targeted over more compact geometries. Long, thin projections facilitated easy removal from the reef matrix, improved handling and storage of samples, and maximised surface area to volume ratio to maximise tissue mass. Once the turf-covered substrate was free from the reef mass, the piece was waved vigorously in the water column to dislodge any fine sediment that might be entrained within the base of the algal matt or within the holfast
83 weave. This procedure reduced the probability that fine sediments would be taken up for analysis when algae was removed from the substrate during laboratory processing.
After collection, turf covered pieces were carried to the surface, immediately rinsed gently with DI water, placed in double zip-seal bags, and placed on ice in complete darkness for transport to the on-island laboratory. At the on-island laboratory algae samples were placed in complete darkness in the refrigerator at 4o C until shipment to the analytical laboratory. To avoid rupture of cells and thus potential loss of Hg from tissue, turf algae samples were not frozen for storage or shipment.
3.6.2 Herbivorous Fish Collection (surgeonfish, Acanthurus lineatus)
Twenty A. lineatus specimens were collected from each marine bay for the Loa,
Vatia and Aasu study sites, consistent with study sites for marine water and turf algae
(Figures 3-4B, 3-5B and 3-7B). In the field, surgeonfish were collected during nighttime dives, within two hours after dusk, by hand spear using SCUBA. All specimens were speared in the head to avoid puncture of muscle tissue or the gut, which could lead to tissue contamination by trace Hg in the environment or through discharge of body fluids. After spearing, each specimen was returned to the boat for euthanisation in accordance with the method approved by the Animal Ethics Committee (UOW
AE07/09, 2007). For euthanisation, fish were immediately placed in an ice bath in a clean polypropylene cooler maintained at ~5o C, and held for 1 min. Fish were then rinsed with DI water, and placed in labeled zip-seal bags on ice for transport to the on- island laboratory.
Zip-seal bags used for fish storage were over-sized so that removal of fish was not required for taking body measurements or for identification. Bagged fish were laid flat on a clean surface, identified to species, measured for standard length and fork length,
84 and weighed. Weight of the bag was accounted for by using an identical bag as tare weight. To determine bag tare weight, five bags from each lot of bags used were weighed to assess variability in bag weight. In no case did bags differ from ± 0.1 g.
Data were recorded and whole fish were frozen at -20o C for storage at the on-island laboratory until shipment to the analytical laboratory.
3.6.3 Carnivorous Fish Collection (goatfish, Mullidae spp.)
Due to availability, Mullidae collections did not target a single species. Goatfish in
American Samoa are represented by 13 species, none of which appear in large schools, nor consistently among all reef habitats (Wass, 1984; P. Brown, US National Park
Service, pers. comm., 2007). Moreover, there is considerable variability among species’ feeding behaviors, with about half the species each in daytime and nighttime feeding regimes (Myers, 1999; Randall, 2005). All Mullid specimens were collected from the same area where sediment samples were collected, i.e., basin sediments below the lower reef margin. A total of 20 Mullidae spp. were collected from each marine bay for the Loa, Vatia and Aasu study sites, consistent with study sites for marine water, turf algae, and herbivorous fish (Figures 3-4B, 3-5B and 3-7B). Mullidae specimens were collected, handled, and stored using procedures described for A. lineatus (see Section
3.6.2).
3.6.4 Carnivorous Fish Collection (barracuda, Sphyraena qenie)
Collections for S. qenie were planned for all coastal waters of Tutuila, seaward from the island to ~1 km, and well away from proximity to the entrance to Pago Pago
Harbour. Collection time was targeted for late-afternoon to dusk, when barracuda are known to disperse from the shelter of the reef to feed (Myers, 1999). Range, time, and
85 method of capture were based on the author’s knowledge and experience of the barracuda catch in Tutuila waters. Minimum numbers of field samples or specific sampling locations were not specified for S. qenie, due to the inherent vagaries of off- shore trolling. A total of 3 S. qenie were taken on trolled lures by the author from small sport-fishing craft along the southeast coast of Tutuila, within 1 km of the Tutuila shore between June 2006 and December 2007. Due to fish size, S. qenie were not stored and shipped as whole fish. After body measurements and weight were taken, a 20 g
(approximate) sample of boneless and skinless muscle tissue was removed from the thickest part of the dorsal muscle, using a clean dissection scalpel. Tissue samples were frozen at -20o C at the on-island laboratroy until shipment to the analytical laboratory.
3.6.5 Analytes, Analytical Methods, and Detection Limits for Hg in Coral Reef Biota
Analyses for THg and MeHg in biological tissue used the same atomic fluorescence techniques used for rainfall and marine water samples, and differed only in sample preparation and the digestion of the tissue matrix for analysis (Table 3-6).
Table 3-6 Methods and detection limits for Hg in coral reef biota Biota Analyte Instrument Method1 Achieved DL2, wet-weight
Turf Algae Total Hg CVAF EPA 1631E 0.182 ng g-1
Methyl Hg CVAF EPA 1630 0.435 ng g-1
Fish Total Hg CVAF EPA 1631E 0.105 - 0.108 ng g-1
Methyl Hg CVAF EPA 1630 0.251 - 0.257 ng g-1
All Lipids Bligh-Dyer 0.05 %
Solids Freeze-dry 0.1 %
1(Bligh and Dyer, 1959; US EPA, 2001a, 2002) 2(Battelle Marine Sciences Laboratory, 2008; Columbia Analytical Services, 2008)
86 Algae samples were received refrigerated at the analytical laboratory and were stored at 4o C in complete darkness until processing. Under trace metals clean room conditions, algae were carefully removed from the substrate by mechanical extraction with stainless steel forceps. Care was taken to avoid forceps contact with substrate. To the extent practicable, turf algae material was extracted from the substrate in a manner that avoided fine sediment entrained within the matt base. A 5.0 g sample of algal tissue was retained from each field sample for analysis.
Whole surgeonfish and goatfish, and a 20 g sample of barracuda muscle tissue, were received frozen at the analytical laboratory and were stored at -20o C until sample processing. For the whole fish, a 5.0 g skinless tissue sample was removed from the thickest part of the fish dorsal muscle for analysis. For the barracuda sample, a 5.0 g sample was removed fro the field sample. Tissue samples were freeze-dried and homogenised to a fine powder in a ball mill prior to digestion.
Analytical methods for tissue were similar to methods for rainfall and marine water
(see Section 3.4) and differed only in digestion of the matrix.
For THg, sample digestion was in accordance with the Appendix for Method 1631E
(US EPA, 2002). A 0.5 g aliquot of freeze-dried tissue was placed in the digestion vessel with HNO3/H2SO4 solution and let stand at room temperature for 4 h. After cold digestion, the sample was heated gradually to 125o C and re-fluxed in the hot acid solution for 3 h, or until the brown color, which indicates organics material remains, was gone. After digestion, the sample was diluted to 22 mL with 0.07M BrCl to destroy any remaining organic material, and to oxidise any dissolved MeHg.
For MeHg, sample digestion was in accordance with procedures described for tissue by Bloom (1989). A 0.5 g aliquot of freeze-dried tissue was placed in the digestion vessel with 5 mL of 25% KOH/methanol solution, shaken vigorously, and let stand at
87 room temperature for 2 h. The sample was then heated in an oven at 65o C for 2 h.
After digestion, sample was diluted to 10 mL with methanol and then brought to a final volume of 22 mL with DI water.
Following digestion, quantification of THg and MeHg in tissue samples followed procedures for rainfall as described in Section 3.4.1.2.
Lipids were extracted from tissue samples according to procedures described by
Bligh and Dyer (1959), a modification of the Folch method (Folch et al., 1957), which is deemed suitable for low-lipid matrices such as marine algae and fish tissues (Iverson et al., 2001).
3.6.6 Calculation of Bio-accumulation Factors
Bio-accumulation factors (BAFs) were calculated along the herbivorous and carnivorous trophic gradients, with marine water as the common reference point. A
BAF for each trophic gradient was calculated for the Aasu, Vatia and Loa sites. The un-impacted sites of Aasu and Vatia were also combined for a BAF calculation.
The herbivorous trophic gradient included:
Marine water » Turf algae » A. lineatus muscle tissue.
The carnivorous trophic gradients included:
Marine water » Mullidae spp. muscle tissue;
Marine water » S. qenie muscle tissue.
The BAF for each trophic step was calculated as a simple quotient, using the step as numerator and the previous step as denominator. Mean concentrations for THg and
MeHg in each matrix were used for BAF calculations.
88 3.6.7 Assessment of Human Health Risks from Consumption of Tutuila Reef
Fish
3.6.7.1 Basis for risk-based consumption limits
Risk-based fish consumption limits for MeHg for surgeonfish and goatfish were prepared in accordance with US EPA guidance (US EPA, 2000). The US EPA guidance is presently the most widely recognized and accepted methodology among world health organisations that are responsible for the evaluation of contaminants in aquatic species exposure pathways. Consumption limits determine the allowable number of fish-meals that can be consumed over a given time period (month), based on the concentration of MeHg found in the target tissue. Methyl Hg is not known to be carcinogenic, so non-cancer health endpoints were assessed using a hazard quotient of 1
(US EPA, 2000). US EPA has established an interim reference dose (RfD) of 1 x 10-4 mg kg-1 d-1 for MeHg based on data on neurological effects in children who had been exposed to Hg in utero (Marsh et al., 1987). This RfD was reviewed by several US- based advisory boards including the Science Advisory Board and the National Academy of Sciences, and it was determined that the RfD is scientifically sound and is supported by human and animal studies (ATSDR, 1999; US EPA, 2000).
3.6.7.2 Calculation of risk-based consumption limits for Methyl Hg
Two equations were used to calculate consumption limits based on non-cancer health endpoints for MeHg. Equation 1 is used to establish the maximum allowable daily rate of ingestion based on the concentration of MeHg in fish muscle tissue. Equation 2 is used to establish number of fish meals allowed for consumption during a given one- month period, so not to exceed the average daily intake limit for MeHg. Consumption
89 limits for MeHg in fish muscle tissue are based on a 0.227 kg (8-ounce) meal size, and a body weight of 70 kg.
Equation 1. Non-cancer health endpoints:
CRlim = (RfD)(BW)/(Cm) where:
-1 CRlim = Consumption Rate, maximum allowable limit (kg d )
RfD = Reference Dose (mg of MeHg per kg of body weight per day)
BW = Body Weight (kg)
-1 Cm = Concentration of MeHg in fish muscle tissue (mg kg wet-weight)
Equation 2. Consumption limits:
Meals per month = (CRlim)(Tap)/(MS) where:
-1 CRlim = Consumption Rate, maximum allowable limit (kg d )
MS = Meal Size (kg fish meal-1)
-1 -1 -1 Tap = Time averaging period (365.25 d y ÷ 12 mo y = 30.44 d mo )
One consumption limit table was prepared for goatfish and one for surgeonfish for each study location where fish were collected (see results in Chapter 4). Tables were prepared based on mean MeHg in muscle tissue of fish from respective sites. If the average concentration of MeHg fell between two consumption limits, the more conservative consumption rate was selected. The definition of "unsafe" fish consumption is a consumption limit of less than one meal every two months, as indicated by “none (< ½)” in the tables.
90 CHAPTER 4 . RESULTS and DISCUSSION
4.1 Mercury in Tutuila Rainfall
Research objectives for this study component were achieved by developing a reliable mean for THg and MeHg concentration in Tutuila rainfall, experimentally derived from repeated measurements from a consistent location over a medium-term sampling period, using rigorous ultra-trace field and analytical procedures. Wet deposition rates for
Tutuila were estimated based on Hg concentration in rainfall, and rainfall distribution.
Results for Hg in Tutuila rainfall were evaluated against experimental values from comparably remote northern hemisphere locations and were used to evaluate current global model predictions for Hg in rainfall, and Hg deposition, for this region of the remote South Pacific. The overall objective, to demonstrate Tutuila’s relative global status for exposure to atmospheric Hg, was achieved.
Results for Hg in Tutuila rainfall (Table 4-1, Figures 4-1 and 4-2) as discussed in the paragraphs below indicate that Hg is distributed extensively at the global scale via atmospheric processes. Similarities for THg in Tutuila precipitation and precipitation from northern hemisphere locations, and findings that global models under-estimate Hg occurrence in remote southern hemisphere regions when compared with experimental data from Tutuila, substantiate concerns for Hg as a global pollutant.
To examine contemporary inter-hemispheric patterns for THg in rainfall among remote areas of the globe using results from Tutuila, only oceanic locations, and locations on windward ocean shores preceded by large expanses of open ocean, were considered for comparisons with Tutuila data. Although there is a fairly large body of relatively recent (past ~10 years) data for Hg in rainfall and Hg deposition in northern hemisphere locations, nearly all data are from continental industrial regions or populated coastal areas of Asia, North America, and Europe (see Section 1.2.1). These
91 areas are highly likely to be strongly influenced by regional Hg emissions. This limits the usefulness of these locations for inter-hemispheric comparisons with the remote southern hemisphere location of Tutuila Island.
In the western hemisphere, coastal or oceanic locations of the North Atlantic were considered unacceptable for comparative evaluation with this study. The likelihood of regional influences from Hg emissions from North America and Europe at North
Atlantic locations is supported by Temme et al. (2003) who compiled inter-hemispheric data for total gaseous Hg (TGM) in the atmosphere over the Atlantic Ocean between
70o S and 70o N latitudes for a six year period (1996-2001), and demonstrated that disproportionately high concentrations of TGM are prevalent between 30o-60o N than for other Atlantic latitudes. It appears this pattern has been consistent for many years.
Slemr et al. (1985), Slemr and Langer (1992), and Slemr et al. (1995) showed similar disparities for TGM over the North Atlantic for 1978-1980, 1990, and 1994, respectively.
In the eastern hemisphere, there appears to be a similar disparity in TGM distribution for northern latitudes. Disproportionately high values of TGM were reported at the
Global Atmospheric Watch Site on An-Myun Island, Korea (36o N, 126o E), and when compared to other data from Asian countries, indicated that eastern shores of northern latitudes of the Asian continent are significantly influenced by regional industrial emissions (Nguyen et al., 2007).
With regard to other northern hemisphere locations, there are no known data for atmospheric Hg or Hg in rainfall for the central or western Eurasian continent, but it is reasonable to assume that environmental releases of Hg from the former Soviet Union and its satellites, and from present-day Russia and former Soviet-bloc nations, would
92 make these regions unsuitable as representative remote global study sites for comparison with Hg in rainfall from the relatively pristine environment of Tutuila.
Geographic locations of the northern hemisphere that were considered potentially suitable as representative of remote global locations with regard to prevailing atmospheric and oceanic currents, with negligible regional influence from industrial emission sources, were limited to locations in the central Pacific Ocean, or Pacific
Ocean windward continental shores. Due caution was also exercised regarding sample size and probable quality of analyses when considering the integrity of available data sources. Under these criteria, there were few suitably remote locations for which experimental data for Hg in rainfall were reported in the literature or available from reliable un-published sources, for comparison with results for hg in Tutuila rainfall.
Contemporary investigations that were selected as suitable for comparisons with Tutuila findings were limited to three sources. Steding and Flegal (2002) reported mean THg of 6.0 ±1.81 ng L-1 (n=17) at the Long Marine Lab (36o N, 125o W) on the coast of
California ~200 km south of San Francisco, USA. Two locations for the USA Mercury
Deposition Network (USDA, 2007) along the northern California coast (MDN station
CA20, 42o N, 125o W) and the northern Washington coast (MDN station WA03, 48o N,
125o W) were also considered suitable for comparison with Tutuila. For station CA20
(2006-2008), mean THg was 3.63 ±0.48 ng L-1 (n=42). For station WA03 (2007) mean
THg was 4.06 ±0.75 ng L-1 (n=29).
Evaluation of the data among these locations showed that data was not normally distributed, which precluded the use of parametric methods for statistical comparisons among sites, although variance for data among locations was shown to be homogeneous. The non-parametric Mann-Whitney U Test was used to test for differences in THg in rainfall between Tutuila and the Long Marine Lab, and between
93 Tutuila and each of the MDN sites. There was no difference in THg concentration in rainfall between Tutuila and the Long Marine Lab (p ~0.95). Concentration for THg in
Tutuila rainfall was significantly higher than for MDN station CA20 (p =0.037) and station WA03 (p =0.011).
These are compelling findings. American Samoa is 8000 km from these northern hemisphere locations. Given the geographical similarities among these sites, and similarities in study scope and sampling periods (1 year or more), there is some basis to conclude that Hg is well mixed and distributed among the Pacific Ocean sectors of the northern and southern hemispheres, although there are factors that influence Hg in precipitation that were not considered here because they were beyond the scope of this study. That Hg deposition in the remote South Pacific does not appear to be significantly different than for precipitation in comparable locations of the northern hemisphere is in need of explanation through further studies.
Results presented here for Hg in Tutuila precipitation, compared with available inter- hemispheric data for direct measurements for Hg in rainfall for remote locations, reinforce accumulated evidence from the past 20 years that Hg is extensively distributed inter-hemispherically via atmospheric processes throughout the world. Neither historic nor contemporary results for Hg in rainfall in remote Pacific locations indicate a significant inter-hemispheric difference for Hg in rainfall during the past two decades.
Historic results for Pacific sectors of both hemispheres are comparable (see Section
1.2.1), and appear generally lower than recent findings for Tutuila, the Long Marine
Lab, and the MDN sites, although not remarkably so. An increase in Hg depositional flux to remote global locations during the past 20 years is, however, indicated.
Numbers of samples for historical investigations are too few for robust statistical comparisons with this study, and even if the earlier works are disregarded because of
94 small sample size or concerns for pre-mid-1990s analytical capabilities, results from this study lead to the same conclusion; there is basis to suggest that there is no inter- hemispheric difference for Hg in rainfall among remote areas across northern and southern reaches of the globe in the contemporary era.
Table 4-1 Hg in Tutuila rainfall; volume-weighted mean Volume THg MeHg1 Month Sample ID (mL) (ng L-1) (ng L-1)
Feb 2007 R1 235 19.3 0.0080 R2/R3 537 4.31 0.0080 Mar 2007 R4/R5 195 16.0 0.0080 R6 190 4.68 0.0080 Apr 2007 R7/R8 425 8.66 0.0491 R9 300 5.58 0.0342 May 2007 R10 400 1.58 0.0094 R11/R12 600 3.72 0.0224 Jun 2007 R13/R14 500 2.36 0.0406 R16 240 4.96 0.0385 Jul 2007 R17/R18 270 4.71 0.0183 R19 520 7.97 0.0094 Aug 2007 R20/R21 470 4.72 0.0316 R22 120 5.09 0.0219 Sep 2007 R23/R24 1000 4.07 0.0230 R25 540 2.46 0.0094 Oct 2007 R26/R27 310 4.61 0.0355 R28 340 5.84 0.0258 Nov 2007 R29/R30 465 6.93 0.0094 R31 275 11.8 0.0094 Dec 2007 R32/R33 380 3.82 0.0499 R34 290 3.27 0.0094 Jan 2008 R35/R36 950 5.41 0.0094 R37 480 4.51 0.0094
Volume-weighted mean: 5.45 ±0.85 0.0204 ±0.0029
1Bold indicates analytical result < DL; value is ½ DL for calculation purposes
95 Figure 4-1 Total Hg in Tutuila rainfall (corrected)
20.0
18.0 Mean (volume-weighted) = 5.45 ±0.85 ng L-1 16.0 (DL = 0.188 ng L-1)
14.0 ) -1 12.0
10.0
8.0 Total(ng Hg L
6.0
4.0
2.0
0.0
R1 R6 R9
R10 R16 R19 R22 R25 R28 R31 R34 R37
R2/R3 R4/R5 R7/R8
R11/R12 R13/R14 R17/R18 R20/R21 R23/R24 R26/R27 R29/R30 R32/R33 R35/R36 Feb '07 Mar '07 Apr '07 May '07 Jun '07 Jul '07 Aug '07 Sep '07 Oct '07 Nov '07 Dec '07 Jan '08 Rainfall Samples (RX/RY indicates mean of paired samples for sampling event)
Fig 4-2 Methyl Hg in Tutuila rainfall (corrected)
0.070
0.060 Mean (volume-weighted) = 0.0204 ±0.0029 ng L-1
0.050 -1
) (DL = 0.0159 and 0.0188 ng L ) -1
0.040
0.030 Methyl(ng Hg L
0.020
0.010
0.000
R1 R6 R9
R10 R16 R19 R22 R25 R28 R31 R34 R37
R2/R3 R4/R5 R7/R8
R11/R12 R13/R14 R17/R18 R20/R21 R23/R24 R26/R27 R29/R30 R32/R33 R35/R36 Feb '07 Mar '07 Apr '07 May '07 Jun '07 Jul '07 Aug '07 Sep '07 Oct '07 Nov '07 Dec '07 Jan '08 Rainfall Samples (RX/RY indicates mean of paired samples for sampling event)
96 Modeling predictions were shown to under-estimate Hg in rainfall and atmospheric wet deposition rates in the remote global location of Tutuila when compared against the experimentally-derived values (Tables 4-1 and 4-2). As discussed in Section 1.1.1, modeling is presently used extensively to support concerns for Hg as a global pollutant, as an important technique used to promote national and international regulatory controls on Hg emissions, and is used principally to compensate for the paucity of experimental measurements of Hg in remote global locations, especially the southern hemisphere.
Table 4-2 Estimated annual Hg wet deposition, Tutuila Island Total Hg Methyl Hg
Catchment (μg m-2 yr-1) (g yr-1) (μg m-2 yr-1) (g yr-1)
Masausi 16.1 25 0.060 0.094
Alega 22.1 29 0.083 0.108
Vatia 23.1 113 0.086 0.423
Tafeu 23.3 119 0.087 0.447
Aasu 26.3 222 0.099 0.832
Tutuila Island 22.6 3085 0.085 11.5
Note: Flux calculated using volume-weighted mean for Hg in Tutuila rainfall (Table 4-1) and rainfall depth polygons from PRISM 2006 rainfall atlas for Tutuila (see Section 3.4.1.5 for flux calculation method).
Experimentally-derived wet deposition rates for THg on Tutuila (~20 μg m-2 yr-1) were twice the model prediction reported by Bergan et al. (1999) who estimated ~10 μg m-2 yr-1 THg for the oceanic region of Tutuila based on their climatological transport model MOGUNTIA. Petersen et al. (1998) showed a similar under-estimate in model predictions for remote locations. Their Tropospheric Chemistry Module of the Eulerian model predicts ~ 2-3 ng L-1 THg in rainfall when atmospheric concentrations of Hg are in the range of ~1 ng m-3, the range measured for the region of Tutuila (Fitzgerald,
1995). Similar to the prediction made by Bergan et al. (1999), the Petersen et al. (1998) model under-estimates THg in rainfall for the remote oceanic region of Tutuila Island
97 by a factor of ~2. The numerical model applied by Shia et al. (1999) does not provide a prediction for Hg in rainfall, but does estimate global Hg depositional flux; for the region of Tutuila they predict 5-10 μg m-2 yr-1, again, an under-estimation by a factor of at least 2 compared with experimental data from Tutuila.
An interesting observation for Hg concentration in precipitation at inter-hemispheric
Pacific Ocean locations is that the TGM concentration gradient described by Fitzgerald
(1995) of ~1 ng m-3 (south) to ~2 ng m-3 (north) does not appear to be reflected in the wet scavenging of Hg from the atmosphere; THg in rainfall is about equal in Pacific northern and southern hemispheres, as shown above. An important implication from this is that the observed gradient may not be relevant in inter-hemispheric Hg wet deposition rates. Thus, models that predict lower Hg deposition rates for remote southern regions of the globe than for northern hemisphere locations, based on measured or predicted atmospheric concentration of TGM, might be expected to under- estimate actual deposition rates for remote southern hemisphere locations.
Inconsistencies between experimentally-derived values and model predictions are usually attributable to an under-estimation of global anthropogenic emissions of Hg, since estimated TGM is a key variable for model inputs (Slemr et al., 2003). As indicated above, the difference in inter-hemispheric atmospheric concentration of TGM may have less relevance for Hg wet deposition than previously thought. Differences in model predictions and experimentally-derived fluxes could also be attributable to model assumptions for a rainfall depth for an oceanic region that is lower than for an island situation. Although data on model sensitivities were not available, and a determination of discrepancies between experimentally-derived flux values and model predictions were beyond the scope of this study, generally favourable agreement between model
98 predictions and experimental data from the remote location of Tutuila Island serve to strengthen model-based conclusions that Hg is a pollutant of global concern.
A potentially complicating factor for these conclusions is that during a 1980 cruise,
Fitzgerald et al. (1984) identified a source of TGM to the equatorial Pacific atmosphere in the region of 4o N - 10o S along the 160o W meridian, that was attributed to deep ocean upwelling. This area is ~1500-2000 km from American Samoa and is in the turbulent region of tradewinds convergence. It is not possible to determine the effects that a relatively localised oceanic source of TGM would have on Hg concentration in rainfall on Tutuila, but this seminal work by Fitzgerald et al. (1984) on the importance of the oceans as a source for atmospheric Hg serves to further support extensive inter- hemispheric mixing and distribution of Hg. More recent work on oceanic Hg processes
(Fitzgerald and Mason, 1996; Mason and Sheu, 2002) show that most Hg transported to the oceans is returned to the atmosphere by evasion. Contrary to this line of thought is that most Hg that enters the ocean surface is removed to the deep ocean on settling particles (~70%) (Lamborg et al., 2002). Either argument supports the importance of oceans in the global cycling of Hg. Residence time of Hg in the ocean surface mixed layer (to 75 m) is relatively short at 4-7 years (Gill and Fitgerald, 1987), and in the ocean overall ~350 years (Gill and Fitgerald, 1988). Given this rapid turnover, and that oceanic circulation is global in extent, it is expected that the oceans play a significant role in the extensive inter-hemispheric distribution of Hg. Fitzgerald (1995) reported
TGM over the central and eastern Pacific between 55o S - 55o N (see Section 1.2.1) and indicated no anomalously high concentration of atmospheric Hg in the Samoa Islands latitudes. Thus, it appears unlikely that a significant bias was introduced for Hg in
Tutuila rainfall because of evasions of TGM from ocean surfaces at the Central Pacific location identified by Fitzgerald et al. (1984).
99 Similarity in deposition rates among study sites and island-wide was expected because wet deposition is rainfall dependent, and most study sites are under the orographic influence of Tutuila’s mountainous interior, and therefore have similar rainfall patterns. The marginal exception is Masausi, which is furthest east and furthest from the central region of the island compared to other study sites, with correspondingly lowest rainfall, and therefore lowest Hg wet deposition rates.
Measured concentrations of Hg in rainfall, and calculated wet deposition rates for Hg on Tutuila Island in the remote South Pacific, strongly support prevailing concerns for
Hg as a global pollutant. Experimental data from Tutuila that corroborates global models for Hg deposition show that models can be reliable (though possibly under- estimates) in supporting assumptions for the proliferation of Hg in the remote global environment.
Results presented in Table 4-1, and Figures 4-1 and 4-2 are final values prepared from a series of manipulations of analytical results. Integrity of data is a major concern for research on Hg, especially for remote regions, because of trace characteristics of Hg in environmental media in remote locations, and the potential for contamination that may confound results. Procedures and manipulations used to prepare the rainfall data to meet research objectives are explained in Section 4.1.1 - 4.1.3, to provide assurances of data integrity.
4.1.1 Acid Blank and Field Blank Correction for Rainfall Samples
Correction factors were applied to laboratory analytical results to account for trace
Hg contamination in acid used for preservation of field samples (Table 4-3) and to account for trace Hg contamination from field sampling equipment or sample handling
(Table 4-4). Correction factors for acid were calculated based on rainfall sample
100 volume, volume of HCl added for sample preservation, and concentration of THg and
MeHg in the acid (negative values indicate subtraction from analytical results). Acid blank correction was not applied to analytical results for MeHg because this species was not detected in the HCl used for sample preservation.
Field blanks were acid-preserved using the same procedures and HCl used to preserve rainfall samples, so an acid correction factor was applied to field blanks, and was accounted for in the overall field blank correction (Table 4-4). For sampling events where replicates were analysed as part of the laboratory QA/QC program (designated r1, r2, see Appendix A), means for replicates were calculated after acid and field blank corrections were applied to analytical results, and were used for the final reported result
(Tables 4-5 and 4-6).
Blank correction of analytical results for Hg in Tutuila rainfall in no way influenced the quality of the determinations presented in the preceding section. For THg, the total correction (acid blank + field blank) was ~1.2 ng L-1, less than 10% of the lowest THg measurement in rainfall (1.28 ng L-1, sample R15).
Table 4-3 Acid blank correction for HCl preservation of Tutuila rainfall samples Total Hg Sample Volume HCl in HCl Correction Month ID (ml) (ml) (ng L-1) (ng L-1)
Feb 2007 R1 235 1.50 6.93 -0.044 R2 537 2.75 6.93 -0.035 R3 537 2.75 6.93 -0.035 Mar 2007 R4 195 1.00 6.93 -0.036 R5 195 1.00 6.93 -0.036 R6 190 1.00 6.93 -0.036 Apr 2007 R7 425 2.20 6.93 -0.036 R8 425 2.20 6.93 -0.036 R9 300 1.50 6.93 -0.035 May 2007 R10 400 2.00 6.93 -0.035 R11 600 3.00 6.93 -0.035 R12 600 3.00 6.93 -0.035
101 Table 4-3 Acid blank correction for HCl preservation of Tutuila rainfall samples (cont’d) Total Hg Sample Volume HCl in HCl Correction -1 -1 Month ID (ml) (ml) (ng L ) (ng L )
Jun 2007 R13 500 2.50 6.93 -0.035 R14 500 2.50 6.93 -0.035 R15 250 1.30 6.93 -0.036 R16 240 1.30 6.93 -0.036 Jul 2007 R17 270 1.35 6.93 -0.035 R18 270 1.35 6.93 -0.035 R19 520 2.65 6.93 -0.035 Aug 2007 R20 470 2.40 6.93 -0.035 R21 470 2.40 6.93 -0.035 R22 120 1.00 6.93 -0.058 Sep 2007 R23 1000 5.00 6.93 -0.035 R24 1000 5.00 6.93 -0.035 R25 540 2.70 6.93 -0.035 Oct 2007 R26 310 1.60 6.93 -0.036 R27 310 1.60 6.93 -0.036 R28 340 1.70 6.93 -0.035 Nov 2007 R29 465 2.35 6.93 -0.035 R30 465 2.35 6.93 -0.035 R31 275 1.40 6.93 -0.035 Dec 2007 R32 380 1.95 6.93 -0.036 R33 380 1.95 6.93 -0.036 R34 290 1.50 6.93 -0.036 Jan 2008 R35 950 4.75 6.93 -0.035 R36 950 4.75 6.93 -0.035 R37 480 2.45 6.93 -0.035
Table 4-4 Field blank (FB) correction for Tutuila rainfall samples Total Hg (ng L-1) Methyl Hg (ng L-1) Vol HCl in in in in FB ID (ml) (ml) FB1 HCl2 FB correction3 FB1 HCl2 FB correction3
FB1 500 2.50 0.480 -0.035 0.445 -0.074 0.008 0 0.008 0
Blank Water 500 2.50 0.406 -0.035 0.371 0.008 0 0.008
FB2 500 2.50 0.192 -0.035 0.157 -0.098 0.044 0 0.044 -0.014
Blank Water 500 2.50 0.094 -0.035 0.059 0.030 0 0.030
Total Hg (mean FB1 & FB2) = -0.086 Methyl Hg (mean FB1 & FB2) = -0.007
1Bold indicates analytical result < DL; value is ½ DL for calculation purposes 2Correction for acid used for FB preservation; see Table 4-3 3Field blank correction = HCl corrected FB - Blank Water
102 Table 4-5 Total Hg in Tutuila rainfall; acid and field blank corrected HCl FB THg THg Data THg correction correction corrected (mean) Month Sample ID Point (ng L-1) (ng L-1) (ng L-1) (ng L-1) (ng L-1)
Feb 2007 R1 R1 19.40 -0.044 -0.086 19.27 19.3 R2-r1 R2 4.71 -0.035 -0.086 4.59 4.71 R2-r2 4.95 -0.035 -0.086 4.83 R3 R3 4.04 -0.035 -0.086 3.92 3.92 Mar 2007 R4-r1 R4 16.10 -0.036 -0.086 15.98 16.0 R4-r2 16.20 -0.036 -0.086 16.08 R5 R5 16.00 -0.036 -0.086 15.88 15.9 R6 R6 4.80 -0.036 -0.086 4.68 4.68 Apr 2007 R7-r1 R7 9.29 -0.036 -0.086 9.17 9.20 R7-r2 9.35 -0.036 -0.086 9.23 R8 R8 8.25 -0.036 -0.086 8.13 8.13 R9 R9 5.70 -0.035 -0.086 5.58 5.58 May 2007 R10 R10 1.70 -0.035 -0.086 1.58 1.58 R11-r1 R11 2.58 -0.035 -0.086 2.46 2.47 R11-r2 2.61 -0.035 -0.086 2.49 R12 R12 5.08 -0.035 -0.086 4.96 4.96 Jun 2007 R13 R13 2.49 -0.035 -0.086 2.37 2.37 R14-r1 R14 2.45 -0.035 -0.086 2.33 2.34 R14-r2 2.48 -0.035 -0.086 2.36 R15 R15 1.40 -0.036 -0.086 1.28 1.28 R16 R16 5.08 -0.036 -0.086 4.96 4.96 Jul 2007 R17 R17 6.38 -0.035 -0.086 6.26 6.26 R18-r1 R18 3.27 -0.035 -0.086 3.15 3.17 R18-r2 3.31 -0.035 -0.086 3.19 R19 R19 8.09 -0.035 -0.086 7.97 7.97 Aug 2007 R20 R20 4.70 -0.035 -0.086 4.58 4.58 R21-r1 R21 4.92 -0.035 -0.086 4.80 4.85 R21-r2 5.03 -0.035 -0.086 4.91 R22 R22 5.23 -0.058 -0.086 5.09 5.09 Sep 2007 R23-r1 R23 3.68 -0.035 -0.086 3.56 3.55 R23-r2 3.66 -0.035 -0.086 3.54 R24 R24 4.71 -0.035 -0.086 4.59 4.59 R25 R25 2.58 -0.035 -0.086 2.46 2.46 Oct 2007 R26 R26 4.03 -0.036 -0.086 3.91 3.91 R27 R27 5.44 -0.036 -0.086 5.32 5.32 R28 R28 5.96 -0.035 -0.086 5.84 5.84 Nov 2007 R29-r1 R29 7.18 -0.035 -0.086 7.06 6.75 R29-r2 6.56 -0.035 -0.086 6.44 R30 R30 7.23 -0.035 -0.086 7.11 7.11 R31 R31 11.90 -0.035 -0.086 11.78 11.8 Dec 2007 R32 R32 3.23 -0.036 -0.086 3.11 3.11 R33-r1 R33 4.51 -0.036 -0.086 4.39 4.52 R33-r2 4.78 -0.036 -0.086 4.66 R34 R34 3.39 -0.036 -0.086 3.27 3.27 Jan 2008 R35-r1 R35 6.16 -0.035 -0.086 6.04 5.98 R35-r2 6.05 -0.035 -0.086 5.93 R36 R36 4.95 -0.035 -0.086 4.83 4.83 R37 R37 4.63 -0.035 -0.086 4.51 4.51
103 Table 4-6 Methyl Hg in Tutuila rainfall; field blank corrected FB MeHg MeHg Data MeHg1 correction corrected1 (mean)1 Month Sample ID Point (ng L-1) (ng L-1) (ng L-1) (ng L-1)
Feb 2007 R1 R1 0.0080 -0.007 0.0080 0.0080 R2-r1 R2 0.0080 -0.007 0.0080 0.0080 R2-r2 0.0080 -0.007 0.0080 R3 R3 0.0080 -0.007 0.0080 0.0080 Mar 2007 R4 R4 0.0080 -0.007 0.0080 0.0080 R5 R5 0.0080 -0.007 0.0080 0.0080 R6-r1 R6 0.0080 -0.007 0.0080 0.0080 R6-r2 0.0080 -0.007 0.0080 Apr 2007 R7 R7 0.0584 -0.007 0.0514 0.0514 R8-r1 R8 0.0425 -0.007 0.0355 0.0468 R8-r2 0.0650 -0.007 0.0580 R9 R9 0.0412 -0.007 0.0342 0.0342 May 2007 R10-r1 R10 0.0094 -0.007 0.0094 0.0094 R10-r2 0.0094 -0.007 0.0094 R11 R11 0.0424 -0.007 0.0354 0.0354 R12 R12 0.0233 -0.007 0.0094 0.0094 Jun 2007 R13-r1 R13 0.0510 -0.007 0.0440 0.0467 R13-r2 0.0564 -0.007 0.0494 R14 R14 0.0404 -0.007 0.0334 0.0344 R15 R15 0.0563 -0.007 0.0493 0.0493 R16 R16 0.0455 -0.007 0.0385 0.0385 Jul 2007 R17 R17 0.0094 -0.007 0.0094 0.0094 R18 R18 0.0342 -0.007 0.0272 0.0272 R19 R19 0.0094 -0.007 0.0094 0.0094 Aug 2007 R20 R20 0.0272 -0.007 0.0202 0.0202 R21-r1 R21 0.0447 -0.007 0.0377 0.0430 R21-r2 0.0552 -0.007 0.0482 R22 R22 0.0289 -0.007 0.0219 0.0219 Sep 2007 R23-r1 R23 0.0319 -0.007 0.0249 0.0255 R23-r2 0.0331 -0.007 0.0261 R24 R24 0.0275 -0.007 0.0205 0.0205 R25 R25 0.0218 -0.007 0.0094 0.0094 Oct 2007 R26 R26 0.0550 -0.007 0.0480 0.0480 R27 R27 0.0300 -0.007 0.0230 0.0230 R28 R28 0.0328 -0.007 0.0258 0.0258 Nov 2007 R29-r1 R29 0.0094 -0.007 0.0094 0.0094 R29-r2 0.0094 -0.007 0.0094 R30 R30 0.0094 -0.007 0.0094 0.0094 R31 R31 0.0226 -0.007 0.0094 0.0094 Dec 2007 R32 R32 0.0430 -0.007 0.0360 0.0360 R33 R33 0.0707 -0.007 0.0637 0.0637 R34 R34 0.0094 -0.007 0.0094 0.0094 Jan 2008 R35-r1 R35 0.0239 -0.007 0.0094 0.0094 R35-r2 0.0094 -0.007 0.0094 R36 R36 0.0094 -0.007 0.0094 0.0094 R37 R37 0.0094 -0.007 0.0094 0.0094
1Bold indicates analytical result < DL; value is ½ DL for calculation purposes
104 4.1.2 Paired Sampling Analysis for Rainfall Samples
Paired samples for THg showed a high degree of variability in relative percent difference (RPD) among paired-sample sampling events (1-50%) although for 9 of 12 paired samples the range was 1-27% (Table 4-7, Figure 4-3). Paired samples for MeHg showed a similar pattern (0-53%) though for 9 of 12 paired samples the range was 0-
26% (Table 4-8, Figure 4-4). An evaluation of results for laboratory QA/QC provided no evidence that variability in paired sample RPD was due to analytical artefacts. Field protocols and procedures for trace elements research were strictly adhered to, and were consistent for each paired sampling event. It is improbable that variability in RPD among paired samples was due to field circumstances. Results for Hg in paired rainfall samples support that there is potential for exceptional variability in Hg deposition at the small spatial scale.
Table 4-7 Relative percent difference (RPD) for Total Hg in paired samples for Tutuila rainfall
Sample pair RPD Sample pair RPD
R2/R3 17% R20/R21 6%
R4/R5 1% R23/R24 23%
R7/R8 12% R26/R27 27%
R11/R12 50% R29/R30 5%
R13/R14 1% R32/R33 31%
R17/R18 49% R35/R36 19%
105 Fig 4-3 Total Hg in Tutuila rainfall (corrected) - paired sampling
20.0
18.0
16.0 (DL = 0.188 ng L-1)
14.0 )
-1 12.0
10.0
8.0 Total(ng Hg L
6.0
4.0
2.0
0.0
R1 R2 R3 R4 R5 R6 R7 R8 R9
R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 Feb '07 Mar '07 Apr '07 May '07 Jun '07 Jul '07 Aug '07 Sep '07 Oct '07 Nov '07 Dec '07 Jan '08 Rainfall Samples (columns side-by-side indicate paired samples for single sampling event)
Table 4-8 Relative percent difference (RPD) for Methyl Hg in paired samples for Tutuila rainfall
Sample pair RPD Sample pair RPD
R2/R3 0% R20/R21 53%
R4/R5 0% R23/R24 20%
R7/R8 9% R26/R27 52%
R11/R12 na1 R29/R30 0%
R13/R14 26% R32/R33 43%
R17/R18 na1 R35/R36 0%
1na = calculation of RPD not applicable because one result < DL
106 Figure 4-4 Methyl Hg in Tutuila rainfall (corrected) - paired sampling
0.070
0.060 (DL = 0.0159 and 0.0188 ng L-1)
0.050
) -1
0.040
0.030 Methyl(ng Hg L
0.020
0.010
0.000
R1 R2 R3 R4 R5 R6 R7 R8 R9
R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 Feb '07 Mar '07 Apr '07 May '07 Jun '07 Jul '07 Aug '07 Sep '07 Oct '07 Nov '07 Dec '07 Jan '08 Rainfall Samples (columns side-by-side indicate paired samples for single sampling event)
Methyl Hg appears to occur at trace levels in Tutuila rainfall, and was only detected near the limits of analytical capabilities. Half (18) of the 37 rainfall samples analysed were below the DL for MeHg. Excluding R15 from the data set (see discussion below), only 7 of 36 results were more than 2 times the DL (R7, R8, R13, R16, R21, R26 and
R33), and only 1 result was more than 3 times the DL (R33).
Samples R1 and R15 require discussion regarding inclusion in the data set. R1 is the highest value for THg and at first may appear anomalous compared with results for most rainfall samples. Although R1 is not exceptional compared with R4, R5 and R31, and may represent a valid concentration for the sampling event, the fact that R1 was the first sampling event for rainfall investigations, and is coincidentally the highest value, leads unavoidably to concerns for validity. Laboratory preparation of equipment or execution of field procedures at the onset of the rainfall sampling program may have adversely impacted results, although a retrospective investigation did not identify a
107 source of error in field or laboratory procedures. The value for R1 could potentially be regarded as an outlier since it meets the criterion of being >3 standard deviations from the mean for THg in rainfall. This does not, however, warrant exclusion of R1 from the data, because the data are not normally distributed, and therefore the definition of outlier is not strictly applicable. In favour of the validity of R1 is that along with R4,
R5, R7, R8 and R31, these are the highest values for the 12-month sampling period, and occur during the non-tradewind months (November-April), which suggests a potential seasonal influence on THg in rainfall. Based on the forgoing, the analytical result for
R1 was accepted as valid and was included in the final data set.
Sample R15 is the lowest value among results for THg in rainfall, and was an unsuccessful attempt to complete a time-series for an un-interrupted rainfall event.
Sample R15 was inadvertently collected without strict adherence to sampling design and field protocols that were followed for all other samples. The rainfall event occurred under adverse field conditions, and was not of sufficient duration to complete the time- series sampling. An investigation into potential sources of error revealed that field events or procedures could potentially have compromised results for R15. These circumstances, combined with the fact that R15 was coincidentally the lowest value for all sampling events, likewise led to questions of validity, and based on these considerations, R15 was excluded from the final data set.
4.1.3 Rainfall for Period January 2007 through June 2008, Alega Village Station
Results for rainfall for the period January 2007 through June 2008 at the Alega
Village gauging station showed that monthly rainfall varied from a low of 124 mm
(June 2007) to a high of 615 mm (March 2007) (Table 4-9, Figure 4-5). Rainfall distribution for the 18-month monitoring period did not show a strong pattern of
108 seasonality, and is consistent with observations for Tutuila that indicate seasonality can be dependent on location (Izuka et al., 2005; US National Climatic Data Center, 2004).
Table 4-9 Monthly rainfall summary for Alega station, Tutuila Island Rainfall Annual totals for successive 12-month periods (mm) Month (mm) 1 2 3 4 5 6 7
Jan '07 543 Feb '07 328 Mar '07 615 Apr '07 162 May '07 532 Jun '07 124 Jul '07 182 Aug '07 136 Sep '07 351 Oct '07 346 Nov '07 377 Dec '07 421 4116 Jan '08 568 4141 Feb '08 214 4027 Mar '08 258 3671 Apr '08 241 3750 May '08 403 3621 Jun '08 222 3719
In general, the non-tradewind months of November-April tended to have increased rainfall compared to the tradewind months of May-October, but this pattern is only roughly demonstrated by the data. As expected, there was considerable inter- and intra- seasonal inconsistency in rainfall patterns, due to the regional influence of the South
Pacific Convergence Zone (see Chapter 1). Inter-seasonal variability for this monitoring period is demonstrated by a relatively dry April 2007 (non-tradewind) which had about the same rainfall as August 2007 (tradewind), and a relatively wet May
2007 (tradewind) which had about the same rainfall as January 2008 (non-tradewind).
Intra-seasonal variability was demonstrated by the disparity in rainfall between March and April 2007, May and June 2007, and January and February 2008.
109 Figure 4-5 Alega station rainfall summary
750 700 Hg in rainfall sampling period 650 period 600 550 500 450 400 350
300 Rainfall(mm) 250 200 150 100 50
0
Jul'07
Sep '07 Sep
Jan'07 Oct '07 Jan'08
Feb '07 Jun '07 Dec '07 Feb '08 Jun '08
Apr Apr '07 Apr '08
Nov '07
Aug '07
Mar '07 Mar '08
May '07 May '08 Month
Measured rainfall at the Alega Village gauging station was used to evaluate rainfall atlas annual predictions for Tutuila (PRISM Group, 2006, Figure 4-6). Annual rainfall totals for 7 successive 12-month periods at Alega varied ±5% from the specified atlas range, in good agreement with atlas predictions (Table 4-10). These results indicated reasonable confidence in atlas predictions, so the atlas data was used to estimate the total annual rainfall volume for each study site catchment, and for Tutuila as a whole.
Table 4-10 Measured rainfall for Alega station vs PRISM atlas prediction 12-month Rainfall PRISM period1 (mm) (range, Alega station, mm) Variation
1 4116 3810 - 3940 +4% 2 4141 3810 - 3940 +5% 3 4027 3810 - 3940 +2% 4 3671 3810 - 3940 -4% 5 3750 3810 - 3940 -2% 6 3621 3810 - 3940 -5% 7 3719 3810 - 3940 -2%
1See Table 4-9
110 Figure 4-6 Rainfall distribution - Tutuila Island (study sites indicated) (from PRISM Group, 2006)
Total rainfall volume, and volume-weighted mean concentrations of THg and MeHg in rainfall, were then used to calculate annual wet deposition rates and total annual deposition to study site catchments and Tutuila overall (see Table 4-2).
4.2 Mercury in Tutuila Marine Water
The principal objective for sampling marine water was to determine a reliable value for Hg in reef waters, to serve as a reference datum for calculation of BAFs (see Section
4.4.3). Correction factors to account for potential contamination from Hg in acid used for sample preservation, or from contaminated sampling equipment, were not required for marine water samples, because there was a negligible addition of THg or MeHg to
111 samples from these sources. For all sites, THg occurred at trace levels, < 4 times the
DL in all samples (Table 4-11). Methyl Hg also occurred at trace levels in marine water, < 5 times the DL for all but one sample at Loa (Table 4-12).
Table 4-11 Total Hg in Tutuila marine water (ng L-1) Study site Sample ID Aasu Vatia Loa
MW1-1 0.378 0.427 0.524 MW1-2 0.374 0.485 0.634 MW1-3 0.389 0.419 0.483 MW1-4 0.484 0.359 0.512 MW2-1 0.278 0.290 0.322 MW2-2 0.282 0.245 0.349 MW2-3 0.286 0.324 0.387 MW2-4 0.254 0.288 0.404 MW3-1 0.295 0.259 0.348 MW3-2 0.271 0.248 0.383 MW3-3 0.391 0.264 0.352 MW3-4 0.232 0.206 0.320
Mean: 0.326 ±0.021 0.318 ±0.025 0.418 ±0.028
Table 4-12 Methyl Hg in Tutuila marine water (ng L-1) Study site Sample ID Aasu1 Vatia1 Loa
MW1-1 0.0094 0.0256 0.0354 MW1-2 0.0094 0.0094 0.0365 MW1-3 0.0239 0.0094 0.0219 MW1-4 0.0265 0.0094 0.0486 MW2-1 0.0359 0.0261 0.0906 MW2-2 0.0347 0.0412 0.0454 MW2-3 0.0414 0.0417 0.0767 MW2-4 0.0346 0.0367 0.0996 MW3-1 0.0094 0.0094 0.0484 MW3-2 0.0262 0.0561 0.0338 MW3-3 0.0244 0.0651 0.0387 MW3-4 0.0232 0.0630 0.0439
Mean: 0.0249 ±0.0032 0.0328 ±0.0061 0.0516 ±0.0070
1Bold indicates analytical result < DL; value is ½ DL for calculation purposes
112 All data for Hg in marine water met assumptions of ANOVA, except for homogeneity of variance for the Aasu MeHg data (p=0.028), and normal distribution for
Loa MeHg data (p=0.043). Given that departures from assumptions were marginal,
ANOVA was used to evaluate differences in THg and MeHg concentrations in marine water among the three study sites. There was a significant difference among sites for
THg (p=0.013) and for MeHg, (p=0.007), attributable in each case to higher Hg levels at Loa, the impacted Harbour site (Figure 4-7). There was no difference in Hg in marine water between the north-shore un-impacted sites of Aasu and Vatia (p>0.05).
Figure 4-7 Hg in Tutuila marine water
0.500
0.400 ) -1 0.300 Total Hg Methyl Hg
0.200 Mean Hg (ng Hg Mean L (DL Total Hg = 0.188 ng L-1)
(DL Methyl Hg = 0.0188 ng L-1) 0.100
0.000 Aasu Vatia Loa
Study site
A non-parametric method was used to confirm differences for Hg among sites because of the relatively small sample size per site (n=12), and the marginal violations of assumptions for ANOVA for MeHg data. In agreement with the parametric test, the
Kruskal-Wallis test showed a significant difference among sites for THg (p=0.026) and
113 for MeHg (p=0.010), attributable to higher Hg levels at Loa, and no difference in marine water Hg between the un-impacted sites of Aasu and Vatia (p>0.05).
Circumstances that may account in part for the slightly elevated levels of Hg in Loa surface waters compared to Aasu and Vatia, are that mixing and exchange of surface waters in the Harbour location are significantly reduced compared to waters of the exposed north shore (US ACOE, 1979).
Although THg and MeHg in reef waters were found in concentrations near the limits of analytical capabilities, the data is considered sound, and comparative analyses should be interpreted with appropriate caution.
When results for Aasu and Vatia are combined (n=24), mean THg was 0.322 ±0.016 ng L-1 (0.206-0.485 ng L-1), and mean MeHg was 0.029 ±0.004 (nd-0.065 ng L-1), in good agreement with historical and recent data for THg in open surface waters across the northern and southern Pacific Ocean. There was no reliable data found for MeHg in open ocean environments from historic or contemporary studies. Compared with historical records, results for THg in Tutuila marine water indicate that Hg concentrations have changed little in near-surface open waters of the North and South
Pacific Ocean over the past 30 years, even though Hg emissions world-wide have increased during this period, and inter-hemispheric distribution of Hg is well supported by results presented in Section 4.1, and by the work of others (see Sections 1.1 and 1.2).
From equatorial and South Pacific cruises of 1980 and 1983, Gill and Fitzgerald (1987) reported THg in marine water from a range of data sets that included: offshore stations in the Tasman Sea beyond the influence of North Island of New Zealand upwelling, with mean THg of 0.56 ±0.08 ng L-1 (0.40-0.72 ng L-1, n=4); along the 160o W meridian between Hawaii and Tahiti, with mean THg of 0.27 ±0.06 ng L-1 (0.17-0.44 ng L-1, n=5); and, also along the 160o W meridian, mean THg of 0.42, 0.44 and 0.40 ng L-1,
114 (for all values, n=4, standard errors and data ranges not given). From a 1990 equatorial
Pacific cruise, Mason and Fitzgerald (1991) reported mean THg in near-surface waters in the range of 0.15-0.35 ng L-1. In the North Pacific, Laurier et al. (2004) reported
THg in surface waters for four cruises in the western, central, and east-northeast sectors from 1980-2002. In the mixed surface layer (~ 75 m), overall mean THg was 0.28
±0.07 ng L-1 for the 1980 cruise, 0.11 ±0.07 ng L-1 for the 1986-1987 cruises, and 0.13
±0.05 ng L-1 for the 2002 cruise. If near-surface samples only are considered (~10 m) for the 2002 cruise, mean THg between Japan and Hawaii ranged from 0.15-0.40 ng L-1, in excellent agreement with results for Tutuila reef waters from a depth of 8 m.
Efficient removal of Hg from the water column was expected to play a major role to limit THg concentrations in Tutuila’s coastal marine water. In the marine mixed surface layer, evasion of volatile Hg species and transport of Hg to deeper layers via sedimentation combine to efficiently remove Hg from the water column. Adsorption of reactive Hg2+ (principal Hg species in rainfall) onto particulate matter, and complexation with organic and inorganic ligands, readily leads to conversion of some dissolved Hg to less soluble forms. A complex transformation pathway involving oxidants and reductants simultaneously leads to production of Hg0 and volatile di- methyl Hg from dissolved Hg2+, and subsequent evasion to the atmosphere from the air- water interface (Nriagu, 1979; Bodek et al., 1988; Fitzgerald and Mason, 1996; Bashkin and Howarth, 2002). The processes of evasion and transport via settling are both important. There is a rapid equilibrium of Hg species in the marine boundary layer, and a relatively short residence time of 4-7 years (Gill and Fitzgerald, 1987).
It is postulated that because of the rapid cycling of Hg in surface waters, the oceans play an insignificant role as a sink for atmospheric deposition of Hg (Fitzgerald and
Mason, 1996) and that a significant build-up of Hg in oceanic surface waters would not
115 be observed over time, even as a result of increases in concentrations of atmospheric Hg by a factor of 3 in the past ~150 years. Fitzgerald and Mason (1996) argue that if anthropogenic Hg emissions ceased today, atmospheric and oceanic conditions for Hg would return to pre-industrial levels within 10-15 years, because the current aerial and oceanic flux of Hg would be deposited and sequestered in terrestrial environments, for gradual release and uptake in coastal and lacustrine aquatic systems. Results for Hg in the marine water column from this study, compared with historic and recent studies for the Pacific Ocean, support these opinions.
Total suspended solids (TSS) were investigated for marine water primarily to provide a qualitative assessment of Hg phase in solution (dissolved or adsorbed) for
Tutuila reefs. Concentrations of TSS appear consistent among study sites (Table 4-13).
Relatively elevated mean TSS for the un-impacted north-shore site of Aasu is likely attributable to the anomalously high readings for MW1-1a and 3a, since all other values for Aasu and other sites lie within the fairly narrow range of ~0.75-1.75 mg L-1. All
MW1 samples for all sites were collected on the same day, within a few hours of each other, and under similar sea and weather conditions. It is possible that a localized event introduced an unaccountable bias in the MW1 samples for Aasu. If the anomalous values are removed from calculation of the mean, the resultant 1.14 mg L-1 for Aasu is in good agreement with the other sites, and the overall mean for all sites is 1.03 mg L-1, which is in excellent agreement with limited historical data for TSS in Tutuila near- shore waters of ~1 mg L-1 (CH2M Hill, 2007; DiDonato, 2007).
In clear waters (<2 mg L-1) the particle-water partition coefficient for Hg is >106, and
Hg would be expected to occur mainly in the <0.45 μm filter-passing “dissolved” phase
(Choe et al., 2003). Partitioning of Hg between dissolved and particulate phases (>0.45
μm) has been quantified by a particle-water partition coefficient (KD) applied to
116 estuarine waters (e.g., Choe et al., 2003) though no application of this method was found for tropical reef waters. Colloid-bound Hg as part of the “dissolved” phase Hg was first identified by Wallace et al. (1982) and Santschi (1982) and may occur as one- half or more of Hg not bound to large particles (>0.45 μm) (Stordal et al., 1996; Choe et al., 2003). For this study, unfiltered reef water samples were analysed for Hg, so a determination of colloid-bound Hg and truly dissolved Hg could not be made. At the range of TSS found for Tutuila reef waters, dissolved Hg (<0.45 μm, filter-passing) is expected to predominate in the reef water column.
Table 4-13 TSS in Tutuila marine water (mg L-1) Study site Sample ID Aasu Vatia Loa
MW1-1a 2.47 0.899 0.700
MW1-3a 2.45 0.816 0.812
MW2-1a 1.57 1.48 0.815
MW2-3a 1.21 0.731 1.74
MW3-1a 0.975 0.885 0.983
MW3-3a 0.802 1.18 1.10
Mean: 1.58 ±0.30 1.00 ±0.11 1.03 ±0.15
Overall mean: 1.20 ±0.13
Of interest was that MeHg occurred in Tutuila near-shore marine water as 7.6%,
10.3 %, and 12.3 % of THg, for Aasu, Vatia, and Loa, respectively, which was ~20-30 times higher than the percentage MeHg reported for Hg in Tutuila rainfall (~0.4%, see
Section 4.1). Enrichment of MeHg as a proportion of Hg in Tutuila near-shore surface waters compared with rainfall input is potentially attributable to a terrestrial source, because significant in situ methylation of Hg in the marine water column is shown to be
117 inconsistent with the biological methylation theory (Weber, 1993; Benoit et al., 1999;
Weber, 1999) because of the relatively high sulfate concentration in seawater (~2700 mg L-1). Another important factor that may further limit methylation in the clear, oligotrophic waters of the tropical coral reef is the availability of appropriate methyl donors, such as large organic molecules (Celo et al., 2006). In the oligotrophic conditions that prevail for tropical coral reef waters, methylation rates might be expected to be limited by the limited availability of methyl donors in the forms of fulvic and humic acids.
Potential limitations on methylation in tropical reef waters, and enriched MeHg in the water column, leads to the suggestion that a sustained terrestrial input (humic-bound
Hg) may be important as a source for MeHg to Tutuila reef systems. As discussed in the following section, elevated Hg in Tutuila reef sediments compared with expected values for marine sediments also suggests a sustained terrestrial input of Hg to Tutuila reefs. Careful further study is needed to determine the relative roles of direct Hg inputs to the water column via stream discharges, or Hg inputs to overlying water via a sediment pathway.
4.3 Mercury in Tutuila Marine Sediments, Stream Suspended Solids, and Upland Soils
Research objectives for the marine sediments study component were met by examining Hg, TOC, and selected elements in sediments for a geographic range of un- impacted open-coastal reef habitats, to examine patterns of Hg distribution and transport among Tutuila reefs, because no such data existed prior to this study. Together, results for Hg and TOC in Tutuila reef sediments (Sections 4.3.1 and 4.3.2), and for Hg and
TOC in stream suspended solids (Section 4.3.3), were used to evaluate the general geographic distribution of Hg on Tutuila un-impacted reefs, for the purpose of selecting
118 representative sites for more detailed studies for Hg in the marine water column
(Section 4.2), bio-accumulation of Hg in marine biota (Section 4.4), and potential human health risks from exposure to the artisanal fishery (Section 4.5).
Of the five un-impacted reef sites, Vatia, with the lowest mean THg concentration in marine sediments and stream suspended solids, and Aasu, with the highest, were selected for study of marine biota, based on these sites occupying the extremes of the range of Hg observed for sediments and stream solids. It was assumed that these sites would provide the greater likelihood of exposing sediment-associated patterns for Hg in water column and biota of un-impacted reef sites, if such patterns existed. The impacted reef habitat of Loa was selected for comparison with the two un-impacted environments to investigate the potential influences from anthropogenically-derived non-point sources of Hg in a remote global environment.
4.3.1. Mercury in Tutuila Marine Sediments
There was a clear progression of increased THg and MeHg in basin sediments from east to west among the five un-impacted sites (Tables 4-14 and 4-15, Figures 4-8 and 4-
9), except for a slight, and insignificant, decrease at Vatia. There was a sharp rise in
THg for the impacted Harbour site of Loa, but only a marginal increase in MeHg at Loa compared with un-impacted sites. Among un-impacted sites, the pattern for Hg in basin sediments is consistent with the pattern for TOC (see Section 4.3.2), and consistent with
THg and TOC concentrations found in stream suspended solids entering each bay (see
Section 4.3.3). The proportion of elements from terrigenous sources in sediments (see
Section 4.3.4) displays a similar pattern among un-impacted study sites.
Depth profiles were provided to show the variability of the terminus of the reef matrix at depth among study sites (Figure 4-10). Stations were coded for graphing
119 purposes, and are referenced to specific study sites (see Chapter 3). Sediment quality, not depth, was the principal criterion for selection of sediment sampling stations on coral reefs. In general, suitable accumulated and stable sediments (hereinafter “basin” sediments) were found in close proximity (within ~2 metres) of the terminus of the reef matrix, except for 3 sampling stations for Vatia, where suitable sediment was found further seaward and at accordingly greater depth.
Table 4-14 Total Hg in Tutuila marine sediments (ng g-1 dry-weight) Study site Sample ID Aasu Tafeu Vatia Loa Alega Masausi 1 13.5 3.74 2.12 21.9 1.84 6.01 2 15.6 3.27 1.58 28.4 3.43 2.59 3 4.93 3.98 2.80 29.9 2.64 2.40 4 15.8 4.31 4.09 40.6 2.03 3.63 5 7.10 3.98 1.86 19.4 4.75 4.25 6 12.9 5.48 1.68 15.1 3.21 3.43 7 7.56 4.54 3.81 29.9 3.12 3.34 8 6.79 5.82 4.34 31.9 3.56 3.03 9 12.2 3.74 3.39 16.0 2.19 2.45 10 3.41 4.95 3.63 21.9 2.00 2.00 11 4.64 5.56 3.21 27.1 5.39 2.58 12 8.46 4.24 3.48 20.7 4.21 2.60 13 10.0 4.00 3.73 25.4 3.37 2.80 14 7.22 3.52 1.82 29.5 2.03 2.76 15 9.29 3.78 1.91 37.7 2.17 4.62 16 7.99 3.37 2.25 39.0 1.54 2.64 17 5.33 5.03 4.70 --- 3.84 3.40 18 7.77 4.98 4.20 --- 3.99 3.97 19 5.99 3.95 5.05 --- 3.34 3.19 20 8.83 3.49 2.54 --- 3.25 3.36 21 5.39 3.36 3.26 --- 5.45 2.52 22 11.4 5.81 3.54 --- 5.17 7.45 23 6.33 3.81 3.16 --- 2.07 4.45 24 7.08 5.48 2.58 --- 4.76 2.61 25 9.44 5.80 2.57 --- 4.80 3.21 26 11.2 5.22 1.32 --- 3.26 2.96 27 6.37 4.20 2.04 --- 4.29 3.51 28 3.93 4.55 2.00 --- 5.30 3.19 29 6.58 5.19 1.90 --- 2.37 4.60 30 9.15 5.18 2.79 --- 5.77 4.01 31 8.00 5.30 3.66 --- 3.60 3.86 32 11.6 4.68 3.46 --- 2.39 5.12
Mean: 8.49 ±0.56 4.51 ±0.14 2.95 ±0.17 27.2 ±1.95 3.47 ±0.22 3.52 ±0.20 Range: (3.41-15.8) (3.27-5.82) (1.32-5.05) (15.1-40.6) (1.54-5.77) (2.00-7.45)
120 Table 4-15 Methyl Hg in Tutuila marine sediments (ng g-1 dry-weight) Study site Sample ID Aasu Tafeu Vatia Loa Alega Masausi 1 0.1822 0.0855 0.0283 0.1790 0.0635 0.0918 2 0.2074 0.0546 0.0265 0.1210 0.1587 0.0294 3 0.0693 0.1327 0.0208 0.2013 0.0795 0.0207 4 0.3725 0.1297 0.0917 0.1702 0.0734 0.1524 5 0.0883 0.0802 0.0491 0.2342 0.1739 0.1184 6 0.2755 0.0846 0.0188 0.1426 0.0646 0.0515 7 0.0815 0.0503 0.0282 0.2685 0.0554 0.0767 8 0.1204 0.1597 0.0491 0.2347 0.1081 0.0331 9 0.3207 0.0658 0.0240 0.1104 0.0536 0.0295 10 0.0321 0.1354 0.1313 0.2326 0.0219 0.0327 11 0.0548 0.0831 0.0462 0.2116 0.1141 0.0467 12 0.1593 0.0840 0.0916 0.1701 0.1264 0.0521 13 0.1664 0.1070 0.0721 0.1371 0.0808 0.1021 14 0.1355 0.0356 0.0191 0.1359 0.0690 0.0270 15 0.1721 0.0723 0.0123 0.1905 0.0407 0.1178 16 0.1446 0.0360 0.0207 0.1712 0.0555 0.0285 17 0.0856 0.1330 0.0635 --- 0.0542 0.0501 18 0.1213 0.1240 0.0568 --- 0.1267 0.1061 19 0.0724 0.0397 0.0285 --- 0.0821 0.0717 20 0.1962 0.0822 0.0586 --- 0.0974 0.0879 21 0.0603 0.0895 0.0710 --- 0.1616 0.0885 22 0.1242 0.2166 0.0417 --- 0.1341 0.1507 23 0.0568 0.0911 0.0475 --- 0.0417 0.1289 24 0.0518 0.1565 0.0684 --- 0.1519 0.0331 25 0.1298 0.1399 0.0572 --- 0.1649 0.0456 26 0.3056 0.0930 0.0534 --- 0.1236 0.0421 27 0.0926 0.0429 0.0315 --- 0.1427 0.0494 28 0.0315 0.0860 0.0534 --- 0.1924 0.1148 29 0.1240 0.1510 0.0586 --- 0.0346 0.0853 30 0.2632 0.1685 0.0552 --- 0.0782 0.0684 31 0.1738 0.2149 0.0826 --- 0.1045 0.0958 32 0.3412 0.1422 0.0866 --- 0.0384 0.1211
Mean: 0.1504 0.1052 0.0514 0.1819 0.0959 0.0734 ±0.0165 ±0.0085 ±0.0047 ±0.0115 ±0.0083 ±0.0069 Range: (0.0315- (0.0356- (0.0123- (0.1104- (0.0219- (0.0207- 0.3725) 0.2166) 0.1313) 0.2685) 0.1924) 0.1524)
The redox horizon in marine sediments (Table 4-16) was documented to validate that sediment samples were collected consistently from oxic conditions among all study sites (sampling depth limited to ~6 cm). Additionally, a distinct oxic/anoxic boundary layer indicated sediment stability, which showed that study design criteria were met for collecting from accumulated and stable sediment deposits that were not subject to significant disturbances by current or wave action.
121
Figure 4-8 Hg in Tutuila marine sediments - all sites (mean w/SE)
35.00
30.00
25.00
20.00 dry-weight)
-1 -1 Total Hg 15.00 Methyl Hg (value)
Hgg (ng 10.00
5.00
0.00 (0.150) (0.105) (0.051) (0.182) (0.096) (0.073) Aasu Tafeu Vatia Loa Alega Masausi Study Site
Figure 4-9 Hg in Tutuila marine sediment - un-impacted sites (mean w/SE)
10.00
9.00
8.00 Total Hg 7.00 Methyl Hg (value) 6.00
dry-weight) 5.00 -1 -1 4.00
3.00 Hgg (ng
2.00
1.00
0.00 (0.150) (0.105) (0.051) (0.096) (0.073) Aasu Tafeu Vatia Alega Masausi Study Site
122 Figure 4-10 Depth profiles for marine sediment sampling stations
0.0 (approx. centerline of bay study site, looking shoreward from sea) Masausi Alega -5.0 Loa Vatia -10.0 Tafeu Aasu
-15.0 Depth(m)
-20.0
-25.0
-30.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Marine Sediment Sampling Station (stations are coded for graphing; see Chapter 3 for locations within bays)
Table 4-16 Redox horizon in vertical profile of Tutuila marine sediments (cm) Study site Measurement Aasu Tafeu Vatia Loa Alega Masausi
1 8.2 6.8 7.3 11.0 6.2 6.2 2 8.0 13.0 7.8 10.0 6.8 9.2 3 7.2 9.5 8.0 11.0 7.7 8.8 4 8.4 7.8 8.0 8.5 13.0 7.0 5 8.2 10.2 8.3 10.5 5.8 7.0 6 7.3 9.8 7.9 9.5 6.1 8.0 7 6.9 8.9 8.2 10.0 5.8 8.0 8 8.1 --- 7.5 8.5 7.9 --- 9 7.9 --- 8.5 8.0 ------10 7.1 --- 7.0 10.5 ------11 ------7.0 ------12 ------8.0 ------13 ------8.5 ------
Mean: 7.7 9.4 7.8 9.8 7.4 7.7
Overall mean: 8.3
123 Statistical comparisons of Hg in sediments among study sites were limited to the impacted (Loa) versus the un-impacted (Aasu and Vatia) sites. Statistical comparisons among un-impacted sites were not attempted because the purpose of this research component was only to establish baseline conditions for distribution of Hg in Tutuila reef sediments, because no previous data was available. Data for Hg in marine sediments did not meet the qualifying assumptions for ANOVA, so the non-parametric
Kruskal-Wallis test was used for comparisons among Loa and the un-impacted sites.
There was a significant difference in sediment THg between Loa and all un- impacted sites (p<0.001). For MeHg, Loa was significantly higher than Masausi,
Alega, Vatia, and Tafeu, (p<0.05), but was not significantly different from Aasu
(p~0.50). Overall, differences in MeHg among sites were marginal, since MeHg concentrations varied only by a factor of about 3 among all sites.
Higher concentration of THg in Loa reef sediments compared to the un-impacted sites, but similar patterns for concentrations of MeHg among all sites, supports the concept that anthropogenically-derived, non-point source Hg contributes to elevated Hg levels at the impacted Harbour location. Elevated THg at the Loa site is associated with only slightly higher MeHg at the Loa site than at the un-impacted sites (generally a factor of ~2-3) as mentioned above. Whereas THg in Loa sediments is about an order of magnitude higher than for THg in sediments at the un-impacted sites, except for
Aasu. Anthopogenically-derived Hg from improper disposal of consumer goods or other waste discharges into the Harbour would presumably contribute elemental Hg to the reef system, on a short-term time scale. Methylation of Hg via the biotic compartments in sediments would presumably not keep pace with elemental Hg input rates, since biotic systems react much more slowly to pollutant inputs than do abiotic compartments (e.g., Harris et al., 2007).
124 Expectedly, there was a fairly strong association between MeHg and THg among un- impacted sites (r2=0.61). Unexpectedly, the mean MeHg:THg ratio among un-impacted sites was 2.1 % (1.7-2.8%), which is somewhat high for marine surficial sediments, which are typically <1 % MeHg (Kannan and Falandysz, 1998; Bloom et al., 1999;
Wasserman, 2002; Sunderland et al., 2006). The ratios found for Tutuila reef sediments are more representative of freshwater systems (Gilmour et al., 1998; Morel et al., 1998).
Elevated ratios for MeHg in Tutuila reef sediments may be corroborating evidence for a sustained and significant terrestrial input of MeHg to Tutuila reef systems, as discussed is Section 4.2, and later in Section 4.3.2 and 4.3.3.
The biogeochemical cycling of Hg in marine systems is extremely complex (e.g.,
Nriagu, 1979; Fitzgerald et al., 2007) and it is beyond the scope or intent of this study to elucidate the patterns discovered for the occurrence and distribution of Hg species in sediments of the tropical reefs of Tutuila Island. Marine sediments have been studied extensively for Hg processes, yet virtually all of this work is for estuarine environments and near-coastal waters of temperate continental regions (e.g., Benoit et al., 1998;
Bloom et al., 1999; Gill et al., 1999; Heyes et al., 2006; Hammerschmidt and Fitzgerald,
2008; Hollweg et al., 2009). Mercury cycling and the processes that control it in marine sediments of tropical coral reefs are simply unknown, and are almost certainly different than for temperate marine sediments from estuaries or coastal margins of continental regions. Factors such as grain size distribution, mineralogy, temperature, organic composition, depth of redox horizon, and species assemblages, undoubtedly create an environment within tropical reef sediments to which known processes for the cycling of
Hg cannot be applied directly. Whether tropical reef sediments contribute MeHg to overlying water and biota in a way similar to the well studied marine sediments can only be determined through careful study.
125 4.3.2 Total Organic Carbon in Tutuila Marine Sediments
The pattern for TOC in marine sediments follows the pattern for THg and MeHg unerringly for the un-impacted sites, but alters for the Loa Harbour site (Table 4-17,
Figure 4-11). Extensive disturbance of the Loa catchment area through engineered fill for structures, construction of impervious pavement, and the distribution of beach sand for landscaping, has almost certainly altered the composition of erosion materials at Loa compared with the un-impacted catchment areas (discussed further in Section 4.3.3).
There was a fairly strong relationship between THg and TOC in basin sediments among un-impacted study sites (r2=0.564), with higher THg generally associated with higher
TOC (Figure 4-12). This pattern is similar for MeHg among un-impacted sites, though less pronounced (r2=0.348) (Figure 4-13).
The importance of organic carbon compounds as receptors of Hg in the terrestrial environment, and as part of the transport and transformation pathways (donors for organic ligands in methylation processes) was introduced in Section 4.2, and will be further discussed in Section 4.3.3 (stream suspended solids) and Section 4.3.4 (upland soils). The patterns for THg and TOC in marine sediments are in excellent agreement with these parameters in stream suspended solids, and are another positive indication that Hg in Tutuila’s aquatic environment may be of significant terrestrial origin.
Table 4-17 Total organic carbon in Tutuila marine sediment (% dry-weight) Study site Sample ID Aasu Tafeu Vatia Loa Masausi Alega 1 0.96 0.29 0.20 0.41 0.32 0.31 2 2.26 0.32 0.25 0.74 0.32 0.51 3 0.55 0.62 0.30 0.43 0.25 0.26 4 3.46 0.47 0.29 1.21 0.49 0.40 5 0.79 0.51 0.34 0.35 0.53 0.40 6 0.87 0.53 0.39 0.28 0.53 0.42 7 1.04 0.32 0.79 0.50 0.57 0.34 8 0.71 0.24 0.30 0.84 0.40 0.43
126 Table 4-17 Total organic carbon in Tutuila marine sediment (% dry-weight) Study site Sample ID Aasu Tafeu Vatia Loa Masausi Alega 9 1.65 0.30 0.25 0.33 0.29 0.34 10 0.56 0.30 0.28 0.27 0.34 0.21 11 1.07 0.39 0.26 0.81 0.38 0.42 12 1.27 0.51 0.29 0.44 0.38 0.31 13 0.86 0.30 0.22 0.43 0.25 0.24 14 2.18 0.31 0.25 0.48 0.33 0.34 15 0.97 0.31 0.18 0.96 0.43 0.38 16 0.85 0.37 0.16 0.96 0.34 0.33 17 0.43 0.44 0.25 --- 0.29 0.35 18 0.67 0.35 0.27 --- 0.43 0.45 19 0.83 0.30 0.35 --- 0.57 0.38 20 0.84 0.24 0.22 --- 0.44 0.33 21 0.57 0.56 0.26 --- 0.42 0.41 22 0.96 0.63 0.23 --- 0.49 0.77 23 0.52 0.28 0.31 --- 0.51 0.46 24 0.77 0.60 0.29 --- 0.28 0.48 25 1.16 0.50 0.33 --- 0.40 0.49 26 0.78 0.41 0.56 --- 0.35 0.45 27 0.64 0.27 0.31 --- 0.38 0.66 28 1.17 0.31 0.28 --- 0.36 0.28 29 2.06 0.46 0.24 --- 0.43 0.38 30 1.14 0.37 0.33 --- 0.31 0.42 31 0.93 0.38 0.42 --- 0.21 0.50 32 1.16 0.56 0.32 --- 0.53 0.34
Mean: 1.08 0.40 0.30 0.59 0.39 0.40
Figure 4-11 Total organic carbon in Tutuila marine sediment (mean w/SE)
1.50
1.00
0.50 TOC (% dry-weight) (% TOC
0.00 Aasu Tafeu Vatia Loa Alega Masausi Study Site
127
Figure 4-12 TOC vs Total Hg in Tutuila marine sediment - un-impacted sites
25.00
20.00 r 2 = 0.564
15.00
dry-weight) -1
10.00
Total (ng Hg g 5.00
0.00 0 1 2 3 4 Total Organic Carbon (% dry-weight)
Figure 4-13 TOC vs Methyl Hg in Tutuila marine sediment - un-impacted sites
0.500
0.400 r 2 = 0.348
0.300
dry-weight) -1
0.200
Methyl (ng Hg g 0.100
0.000 0 1 2 3 4 Total Organic Carbon (% dry-weight)
128 4.3.3 Mercury and Total Organic Carbon in Stream Suspended Solids
There was a generally concomitant increase in THg with TOC in stream solids from east to west for the un-impacted sites of Masausi, Alega, Vatia, Tafeu and Aasu
(Figures 4-14 and 4-15). This pattern agrees closely with rainfall deposition rates established for Hg (see Table 4-2), and closely with the pattern displayed for THg and
TOC in marine sediments from the un-impacted sites (see Figure 4-9 and 4-11).
Together, these patterns strongly suggest climatic and geographical controls on inputs of Hg to reef environments of Tutuila, including Hg concentration in rainfall, rainfall distribution, and catchment area. Combined with results for Hg in Tutuila soils (Section
4.3.4), the overall indication is that rainfall is a likely principal source of Hg in the un- impacted aquatic systems of Tutuila.
Figure 4-14 Total Hg in Tutuila stream suspended solids
225
200
175
150
125
dry-weight) -1 100
Hgg (ng 75
(DL = 1.48 - 3.00 ng g-1) 50
25
0 Aasu Tafeu Vatia Loa Alega Masausi Study site
129 Figure 4-15 TOC in Tutuila stream suspended solids
20
15
10
5 TotalOrganic Carbon dry-weight) (%
0 Aasu Tafeu Vatia Loa Alega Masausi Study Site
Table 4-18 DI water blank correction for Tutuila stream suspended solids Buckets per Sample DI water DI water DI water THg added sample volume source1 used THg1 to matrix2 Sample ID (5 gal) (L) (1,2 or 3) (L) (ng L-1) (ng)
Masuasi S1 3 ~ 57 1 1.15 1.51 1.74 Masausi S2 5 ~ 95 1 0.75 1.51 1.13 Masausi S3 5 ~ 95 3 0.90 0.312 0.281 Alega S1 3 ~ 57 1 1.10 1.51 1.66 Alega S2 5 ~ 95 2 0.75 0.825 0.619 Alega S3 5 ~ 95 3 1.25 0.312 0.390 Loa S1 3 ~ 57 3 0.85 0.312 0.265 Loa S2 5 ~ 95 3 1.10 0.312 0.343 Loa S3 5 ~ 95 2 0.75 0.825 0.619 Vatia S1 3 ~ 57 1 1.20 1.51 1.81 Vatia S2 5 ~ 95 1 0.75 1.51 1.13 Vatia S3 5 ~ 95 3 0.90 0.312 0.281 Tafeu S1 3 ~ 57 1 1.00 1.51 1.51 Tafeu S2 5 ~ 95 3 2.80 0.312 0.874 Tafeu S3 5 ~ 95 3 1.40 0.312 0.437 Aasu S1 3 ~ 57 1 1.05 1.51 1.59 Aasu S2 5 ~ 95 3 2.40 0.312 0.749 Aasu S3 5 ~ 95 2 1.40 0.825 1.16
1Total Hg in DI water source: 1 = 1.51 ng L-1; 2 = 0.825 ng L-1; 3 = 0.312 ng L-1 2For blank correction, assumed all Hg in DI water retained in matrix
130 Correction factors, applied to laboratory analytical results to account for trace Hg contamination in DI water that was used to wash settled solids from sample collection containers, are presented in Table 4-18. Correction factors were calculated in mass units (ng) based on volume of DI water used to process each sample, and concentration of THg in the DI water. For correction purposes, it was assumed that all THg in the DI water was ultimately retained in the dried solids matrix, and was therefore reflected in the final analytical result. This was the most conservative assumption, since it was not practical to determine the actual amount of THg from DI water retained in the sample matrix. Information on QA/QC for THg in stream suspended solids is provided to support data integrity, similar to QA/QC assurances provided for Hg in rainfall.
Table 4-19 Total Hg in Tutuila stream suspended solids; blank corrected (dry- weight) DI water THg THg Dry mass THg correction1 corrected2 (mean)3 Study site Sample ID (g) (ng g-1) (ng) (ng g-1) (ng g-1)
Masausi Masausi S1-r1 1.28 104 1.74 103 101 Masausi S1-r2 1.28 101 1.74 100 Masausi S2-r1 1.15 126 1.13 125 127 Masausi S2-r2 1.15 130 1.13 129 Masausi S3 1.19 106 0.281 106 106 Alega Alega S1 2.87 118 1.66 117 117 Alega S2 7.16 120 0.619 120 120 Alega S3-r1 1.22 132 0.390 132 130 Alega S3-r2 1.22 129 0.390 129 Loa Loa S1 2.83 45 0.265 45 45 Loa S2 7.09 110 0.343 110 110 Loa S3 9.06 97 0.619 97 97 Vatia Vatia S1 0.46 178 1.81 174 174 Vatia S2 0.385 180 1.13 177 177 Vatia S3 0.660 111 0.281 111 111 Tafeu Tafeu S1 0.290 181 1.51 176 176 Tafeu S2 0.286 194 0.874 191 191 Tafeu S3 0.194 146 0.437 144 144 Aasu Aasu S1 0.378 176 1.59 172 172 Aasu S2 0.204 216 0.749 212 212 Aasu S3 0.134 200 1.16 191 191
1See Table 4-18. 2Corrected value based on dry mass of sample; assumed all THg in DI water retained in matrix. 3Applies to replicates only.
131 The impacted Harbour site of Loa had unexpectedly the lowest THg and TOC in stream suspended solids among all sites (Tables 4-19 and 4-20, Figures 4-14 and 4-15), and is likely a paradoxical benefit of anthropogenic development in the Pago Pago
Harbour catchment. Most of the catchment area for the Loa reef is under residential development (GDC, 2007; P. Peshut, pers. obs., 2001-2008), and has been since the
1960s (American Samoa Department of Commerce, unpublished data, 2008). As a result, much of the organic material in surface soils has been removed or covered over.
Over the decades, native soils in stream catchment areas have been extensively removed and replaced by quarried fill for construction of structures on steep slopes. Concrete and asphalt pavement are extensively used throughout the catchment area to reduce erosion and land slippage, and to provide reliable access to dwellings and community structures. Beach sand is a traditional landscaping material for Samoan residences, and is viewed as a communal resource by villagers (Meade, 1928; P. Peshut, pers. obs.,
2001- 2008). Large quantities of sand are regularly removed from the shoreline to upland residences within the catchment area (P. Peshut, pers. obs., 2001-2008). Based on these development aspects, there is a high likelihood that a greater proportion of inorganic material is present in the suspended solids of stream water at Loa than for streams of the un-impacted sites.
Table 4-20 Total organic carbon (TOC) in Tutuila stream suspended solids TOC (% dry-weight) Study site S1 S2 S3 mean
Masausi 5.99 4.90 6.99 6.0 Alega 8.94 8.55 9.71 9.1 Loa 2.54 5.49 5.60 4.5 Vatia 11.4 10.2 9.18 10.3 Tafeu 9.54 8.75 11.7 10.0 Aasu 11.0 10.4 16.0 12.5
132 The pattern that emerged from results for THg and TOC in stream solids has important implications for suggesting a principal source of Hg to Tutuila reef systems, and supports the concept that Hg is transported to Tutuila reef systems via a terrestrial route, as first suggested by the findings of an enriched MeHg:THg ratio in marine water
(see Section 4.2). Regression analysis (Figure 4-16) shows a fairly strong relationship between terrestrial-derived organic matter and the transport of Hg to Tutuila reef systems (r2~0.60). Humic materials are the principal form of organic matter in soils that complex with Hg (Wang et al., 1997; Wang et al., 2003) and even small amounts
(as low as 0.1%) can have a significant effect for retention of Hg within in the soil matrix (Mauclair et al., 2008). Lower THg and TOC overall for Loa compared with un- impacted sites also supports the premise that the organic-rich native soil matrix of
Tutuila is a potentially major source for Hg transported to reef waters, if assumptions for reduced organic matter in runoff within the Loa catchment area are accepted.
Figure 4-16 Hg vs TOC in Tutuila stream suspended solids
225
200
175 r 2 = 0.595
150
125
dry-weight) -1
100
75 Total(ng Hg g
50
25
0 0 5 10 15 20 Total Organic Carbon (% dry-weight)
133 4.3.4 Mercury in Tutuila and Aunu’u Upland Soils
Investigating THg in Tutuila upland soils was intended to provide corroborative information to support whether or not wet deposition of Hg via rainfall is the likely predominant source of Hg to the Tutuila aquatic environment, via a terrestrial pathway.
Original research and periodic published reviews from the extensive peer-reviewed literature on Hg atmospheric processes (see Sections 1.1 and 1.2) show that research continues to support long-range transport and distribution of Hg via the atmosphere, and wet deposition, as the principal source of Hg to aquatic systems, world-wide. Localised geologic sources of Hg are also identified as potentially important sources for Hg accumulation in biota, and therefore should be accounted for to the extent practicable when investigating Hg in remote locations (Rasmussen, 1994; Downs et al., 1998).
Geologic considerations notwithstanding, the preponderance of evidence remains in favour of atmospheric deposition as the major factor in bioaccumulation of Hg, even for remote global environments far from emission sources (e.g., Fitzgerald et al., 1998), with soils acting as a receptor and temporary sink of atmospherically deposited Hg before transformation and subsequent release to aquatic systems (Gustin et al., 2008).
Soil sampling stations for Aunu’u Island (n=9) located ~1.3 km off the southeast coast of Tutuila are not shown, because a graphical representation of Aunu’u was not available at the time of this writing. Results from the Aunu’u samples are included in the table of results, and were included in the overall evaluation of data.
Land-use categories considered to be suitably un-impacted for purposes of investigating background soil Hg in this remote global environment were limited to rural and remote locations, classified according to reconnaissance of sampling areas, and extensive observations of population distribution and anthropogenic impacts throughout Tutuila (P. Peshut, pers. obs., 2001-2006). Tutuila regional designations
134 (NW, etc.) assigned to each sampling location were referenced to the main port facilities at Pago Pago Harbour. Results for Hg in upland soils were evaluated on a regional basis to account for potential differences in the geologic base, in consideration of age-relation and proximity of eruption centres as described by Daly (1924), Stearns
(1944) and Macdonald (1944), as discussed in Chapter 1.
Table 4-21 Hg in Tutuila upland soils (mg kg-1 dry-weight) Land use Total Sample ID1 Catchment Region Topography Category Hg DQ2
AS AATA SO001 Aasu NW Mountainous Remote 0.048 B AS AATA SO002 Aasu NW Coastal fringe Remote 0.005 nd AS AATA SED001 Aasu NW Coastal fringe Remote 0.006 nd AS SITA SED001 Aasu NW Mountainous Remote 0.007 nd AS SITA SED002 Aasu NW Mountainous Remote 0.022 B AS AOTA SED001 Aoloau Sasae NW Mountainous Remote 0.073 B AS AOTA SED002 Aoloau Sasae NW Mountainous Remote 0.140 AS AOTA SED003 Aoloau Sasae NW Coastal fringe Remote 0.005 nd AS TCPTA SO001 Aoloau Sasae NW Mountainous Remote 0.011 nd AS TCPTA SED001 Aoloau Sasae NW Mountainous Remote 0.150 AS TCPTA SED002 Aoloau Sasae NW Mountainous Remote 0.190 AS TCPTA SED003 Aoloau Sasae NW Mountainous Remote 0.092 AS TCPTA SED004 Aoloau Sasae NW Mountainous Remote 0.072 AS UBTA SO001 Auasi SE Coastal fringe Rural 0.087 AS UBTA SO002 Auasi SE Coastal fringe Rural 0.005 nd AS UBTA SO003 Auasi SE Coastal fringe Rural 0.005 nd AS ANBTA SO002 Aunuu Sasae SE Mountainous Rural 0.090 AS ANBTA SO001QA Aunuu Sasae SE Mountainous Rural 0.086 AS AUNUU SO001QA Aunuu Sasae SE Mountainous Rural 0.056 AS CATA SO001 Aunuu Sasae SE Mountainous Rural 0.130 AS CATA SO002 Aunuu Sasae SE Mountainous Rural 0.048 B AS CATA SO003 Aunuu Sasae SE Mountainous Rural 0.029 B AS CATA SO004QC Aunuu Sasae SE Mountainous Rural 0.026 B AS CATA SED001 Aunuu Sasae SE Mountainous Rural 0.078 AS CATA SED002 Aunuu Sasae SE Mountainous Rural 0.150 AS TRTA SO001 Fagaalu SW Mountainous Rural 0.080 AS TRTA SO002 Fagaalu SW Mountainous Rural 0.067 AS FPTA SO001 Fagalii NW Mountainous Remote 0.038 B AS FPTA SO002 Fagalii NW Mountainous Remote 0.030 B AS FPTA SO003 Fagalii NW Mountainous Remote 0.088 AS LPTA SO001 Fagatele Larson SW Coastal plain Rural 0.089 AS LPTA SO002 Fagatele Larson SW Coastal plain Rural 0.064 B AS LPTA SED001 Fagatele Larson SW Coastal plain Rural 0.014 B AS PTA SO001 Fagatele Larson SW Coastal plain Rural 0.101 C AS PTA SO002 Fagatele Larson SW Coastal plain Rural 0.161 C AS PTA SED001 Fagatele Larson SW Coastal plain Rural 0.101 C AS APTA SO001 Leone SW Coastal plain Rural 0.059 C
135 Table 4-21 Hg in Tutuila upland soils (mg kg-1 dry-weight) Catchmen Land use Total Sample ID1 t Region Topography Category Hg DQ2
AS APTA SED001 Leone SW Coastal plain Rural 0.093 B AS APTA SED002 Leone SW Coastal plain Rural 0.087 B AS MATA SO001 Leone SW Coastal plain Rural 0.092 AS MATA SO002 Leone SW Coastal plain Rural 0.076 C AS MATA SED001 Leone SW Coastal plain Rural 0.101 C AS MATA SED002 Leone SW Coastal plain Rural 0.065 C AS MTA SO001 Maloata NW Mountainous Remote 0.070 B AS MTA SO002 Maloata NW Mountainous Remote 0.037 B AS MTA SO004 Maloata NW Mountainous Remote 0.049 B AS MTA SO003 Maloata NW Coastal fringe Rural 0.028 B AS MVMTA SO001 Masefau NE Coastal fringe Remote 0.006 nd AS MVMTA SO002 Masefau NE Coastal fringe Remote 0.061 B AS NITA SO001 Masefau NE Mountainous Remote 0.072 B AS ATTA SO001 Nua Seetaga SW Mountainous Remote 0.081 B AS ATTA SED001 Nua Seetaga SW Mountainous Remote 0.054 B AS ATTA SED002 Nua Seetaga SW Mountainous Remote 0.068 B AS SETA SED001 Nua Seetaga SW Mountainous Remote 0.050 B AS AMTA SO001 Pago Pago SE Mountainous Remote 0.034 B AS AMTA SO002 Pago Pago SE Mountainous Remote 0.077 AS AMTA SED001 Pago Pago SE Mountainous Remote 0.025 B AS AMTA SED002QC Pago Pago SE Mountainous Remote 0.029 B AS MPTA SO001 Tafuna SW Mountainous Rural 0.007 nd AS MPTA SO002 Tafuna SW Mountainous Rural 0.005 nd AS MPTA SO003 Tafuna SW Mountainous Rural 0.011 B AS MPTA SED001 Tafuna SW Mountainous Rural 0.005 nd
1SO=sample collected from dry non-drainage area; SED=sample collected from dry drainage area 2Data Qualifier: nd=not detected at method DL, value is ½ DL; B=value is between method DL and reporting limits; C=value is blank corrected
Levels of Hg found in Tutuila soils indicate that the geologic base of predominantly volcanic olivine basalts is not likely to be a significant contributing source for Hg in Tutuila upland soils (Table 4-21). The mean for THg found in upland soils (0.061 ±0.006 mg kg-1) is markedly higher than the range of Hg reported for basalts in the literature (≤0.010-.0.040 mg kg-1, Section 1.2.2). Moreover, basalts of oceanic origin are shown to be near the lower end of the documented range.
Basaltic rocks of Tutuila’s geologic base appear doubtful as a significant source of soil THg, unless Tutuila olivine basalts are exceptionally high in Hg compared with the reported world-wide range. In the absence of an external Hg source, natural processes
136 of weathering and de-gassing would be expected to reduce Hg in native soils compared with Hg in parent material (Ure and Berrow, 1982). Levels of Hg found in Tutuila soils are contrary to the expected pattern, and indicate a sustained Hg contribution from an external source. On remote Tutuila, rainfall is the only external source of THg that could result in the apparently elevated concentration of Hg in soils compared with native basalts. These observations, combined with evidence for associations of THg and TOC, in stream suspended solids and in reef sediments, strongly support the concept that rainfall is the principal source of Hg in un-impacted aquatic systems of the remote Tutuila environment.
4.3.5 Selected Elements (Other than Hg) in Marine Sediments
The objective for investigating additional elements (other than Hg) in marine sediments was to evaluate associations of non-trace elements with Hg in sediments, to determine if a basis for a sediment quality assessment for Hg could be established with respect to one or more elements. If THg or MeHg were found to be closely correlated with one or more of the common elements, such as Al, Fe, Mn, K, or Si, these elements could serve as indicators of potential Hg contamination in marine sediment. For
Tutuila, evaluation of correlations between common elements and Hg compounds in reef sediments of impacted and un-impacted locations was a first attempt to determine if sediment quality guidelines development for American Samoa was practical. This effort was viewed as a collateral opportunity because of extensive sediment sampling to meet other study objectives.
The overall motivation for this objective was economy, because field collection protocols and laboratory analyses for the common elements are relatively simple and
137 inexpensive compared with those for trace concentrations of Hg, as typically found in environments that are not impacted by an Hg emission source.
There appeared to be no suitably significant associations between the more common elements and Hg in Tutuila reef sediments to justify an Hg sediment quality assessment based on concentrations of common elements. The Spearman rank correlation matrix
(Table C-2, Appendix C) shows robust correlations were found between typically abundant elements in tropical soils of volcanic origin, such as Al and Fe (ρ=0.9588), Al and Ti (ρ=0.9287), Al and Si (ρ=0.9748), Fe and K (ρ=0.9145), Fe and Ti (ρ=0.9763), and Fe and Si (ρ=0.9696). There was a strong inverse relationship between terrestrial elements and dominant marine elements, such as Ca and Al (ρ= -0.9035), Ca and Mn
(ρ= -0.9284) and Ca and Si (ρ= -0.8759). Mercury compounds were generally only weakly correlated with both terrestrial- and marine-derived elements (ρ generally < 0.5), and positive determinations could not be made.
A pattern of element distribution was discernible in the data that reflected the pattern of Hg and TOC distribution observed for marine sediment (Sections 4.3.1 and
4.3.2), and Hg and TOC observed for stream suspended solids (Section 4.3.3), (Table
C-1, Appendix C). There was an obvious general trend of increased Al, Fe, Mn, Ti, Si and As east to west across Tutuila from Masausi to Aasu, which suggests a pattern of increasing terrestrial input, and which is coincidental with rainfall distribution and increased catchment area. There was a concomitant decrease in Ca in the same geographic direction, though the magnitude was less pronounced for this element, where there is a significant marine-derived source compared with the terrestrial elements (Ca can be up to 10% by weight in basalts). Potassium also showed an east to west increase trend, though weaker than for the elements given above. Magnesium and
Sr showed little difference among sites, as would be expected for these elements
138 commonly found to have terrestrial and marine environment sources. No attempt was made to refine observations further, as this would be outside the general scope and intent of this study. Overall, patterns support that terrestrial inputs are an important factor in the distribution of elements other than Hg on Tutuila reefs.
4.4 Mercury in Tutuila Coral Reef Biota
Study sites for investigations of Hg in Tutuila coral reef biota were selected based on the geographic patterns for Hg and TOC in reef sediments, and stream inputs to reefs, as described in the previous sections. North-shore Tutuila sites, Aasu and Vatia, represented pristine “un-impacted” reefs in the remote tropical South Pacific. The Loa site represented a remote coral reef subject to anthropogenic disturbances, but not subject to contamination from any known Hg emission sources.
4.4.1 Mercury in Turf Algae
The objective of this study component was to quantify Hg species in this exclusive food source for the herbivorous surgeonfish Acanthurus lineatus, to calculate Hg accumulation factors for turf algae referenced to water column, to provide intermediate resolution of BAFs for the herbivorous trophic gradient on Tutuila reefs.
Mercury in turf algae varied little among the un-impacted north-shore sites and the impacted Harbour site (Tables 4-22 and 4-23, Figure 4-17), with similar concentrations for Aasu and Loa turf, and only slightly lower levels for Vatia. For THg, all data met the qualifying assumptions for ANOVA, which showed there was no significant difference in turf algae THg among study sites (p>0.05). For Vatia, turf algae MeHg was lower than the other sites, and was below the DL for all samples. Data for MeHg in turf algae did not meet assumptions for ANOVA, so the non-parametric Kruskal-Wallis
139 Test was used to test for differences among sites. Although all sites were significantly different from each other for MeHg, (p<0.001), it was recognised that mean MeHg values differed only by a factor of ~2-5 among sites, and were near the limits of analytical capabilities. For MeHg for all sites, 13 of 30 samples were < DL, 8 of 30 samples were < 2 times the DL, and all but 2 samples were < 3 times the DL, which precludes a robust interpretation of differences in MeHg in turf algae among reef sites.
Table 4-22 Total Hg in Tutuila turf algae (ng g-1 wet-weight) Study Site Sample ID Aasu Vatia Loa
TA-1 9.05 8.09 20.3 TA-2 12.1 10.2 11.7 TA-3 13.4 11.9 19.4 TA-4 12.7 15.5 11.0 TA-5 12.0 6.72 16.1 TA-6 12.5 9.06 13.3 TA-7 16.4 7.76 8.50 TA-8 20.1 12.0 14.3 TA-9 19.5 8.27 9.37 TA-10 12.1 14.8 6.47
Mean: 14.0 ±1.12 10.4 ±0.96 13.0 ±1.44
Table 4-23 Methyl Hg in Tutuila turf algae (ng g-1 wet-weight) Study Site Sample ID Aasu Vatia1 Loa1
TA-1 0.801 0.217 0.656 TA-2 0.972 0.217 0.461 TA-3 0.893 0.217 1.20 TA-4 0.870 0.217 0.217 TA-5 0.971 0.217 0.217 TA-6 1.34 0.217 0.464 TA-7 1.14 0.217 0.217 TA-8 1.48 0.217 0.590 TA-9 0.792 0.217 0.440 TA-10 0.923 0.217 0.471
Mean: 1.02 ±0.07 0.22 0.49 ±0.09
1Bold indicates analytical result < DL; value is ½ DL for calculation purposes
140 Figure 4-17 Total Hg and Methyl Hg in Tutuila turf algae (mean w/SE)
20.00
15.00
10.00
wet-weight) Total Hg -1 Methyl Hg
-1
Hgg (ng (DL Total Hg = 0.182 ng g ) 5.00 (DL Methyl Hg = 0.435 ng g-1)
0.00 Aasu Vatia Loa Study Site
Of considerable interest regarding Hg movement through environmental compartments on remote Tutuila, was the unexpected magnitude of accumulation found between the water column and turf algal tissue. Accumulation factors for THg for marine water » turf algae were similar for all sites, at 104.6, 104.5 and 104.5 for Aasu,
Vatia, and Loa, respectively (Table 4-27, Section 4.4.3). Accumulation factors for
MeHg were marginally more variable than for THg (Table 4-28, Section 4.4.3). For
Aasu, the MeHg accumulation factor was 104.6, the same as for THg. The MeHg accumulation factor for Vatia, 103.8, was an estimated value because all samples were <
DL for MeHg. If the DL is used for calculations rather than one-half the DL as typically used by convention, the accumulation factor for Vatia is 104.3, which is in close agreement with the THg accumulation factor, and shows a pattern similar to THg and MeHg for Aasu. At Loa, the MeHg accumulation factor of 104.0 is also an estimate because one-half the DL is used (3 of 10 samples). If the DL is used, the factor for
MeHg in Loa turf algae is 104.1, in good agreement with the factor for THg, and follows the same pattern as the other two sites. Using the DL rather than one-half DL increases
141 the BAF, because the tissue value for Hg serves as the numerator in the BAF quotient
(see Chapter 3).
It will be shown later that the principal bioaccumulation step for Hg in the herbivorous trophic uptake pathway investigated here is from the water column to turf algae tissue, and that for both THg and MeHg a less significant (10±1) step occurs between turf algae and muscle tissue of the herbivorous Acanthurus lineatus.
The pattern of accumulation between water column Hg and algal tissue Hg, and between algae and fish tissue Hg, is fundamentally similar for Tutuila reefs compared with the pattern described by Mason et al. (1995) for bio-accumulation steps for water, phytoplankton, and fish, where the greatest accumulation (~105) takes place between the abiotic and biotic compartments, via phytoplankton primary production. Ionic Hg species prevail in marine water (Bloom and Crecelius, 1983; Gill and Fitzgerald, 1985;
Gill and Fitzgerald, 1987; Fitzgerald and Mason, 1996), and given the large accumulation factors between water and algal tissue found here, it is reasonable to suppose that dynamics of active transport of Hg across cellular membranes in coral reef turfs are similar to those shown for phytoplankton and macroalgae (Morris and Bale,
1975; Eide et al., 1980; Mason et al., 1995; Mason et al., 1996).
Accumulation factors of >104 for Hg in Tutuila turf algae are noteworthy when fundamental characteristics of turf algae ecology are considered. As the source of most primary production on the reef (80%, see Section 2.3), turf algae are grazed extensively by a great variety of marine herbivores at the micro-, meso- and macro-scales, with biomass turnover ~4-12 days (Carpenter, 1985; Klumpp et al., 1987). Magnification of
Hg in turf algae of Tutuila reefs at the levels found here, consistent across three widely separated reef systems, appear remarkable given the ephemeral nature of turf algae on the reef, and indicate a highly efficient mechanism for Hg uptake from reef waters via
142 the primary production pathway on coral reefs. These findings challenge conventional notions that sediments are the principal sink for Hg pollution, and the principal source for Hg uptake in aquatic food webs.
On coral reefs, turf algae and accumulated basin sediments are distinct habitats, and likely form independent bases for feeding regimes. On Tutuila coral reefs, Hg may enter the food web independent of sediment-associated pathways through uptake by biota that feed on turf algae. The convergence of sediment-based and turf-based uptake pathways in top-trophic predators that feed directly or indirectly on prey from both feeding regimes, could reasonably be expected to be manifested in an increased body burden of Hg, above what might be expected if only sediment-associated uptake pathways were considered. This may help explain the unexpectedly high levels of Hg found in the top-tropic reef predator Sphyraena qenie, even though these fish were of relatively small size and occupied pristine reef environments.
Efficiency of Hg uptake in the turf algae pathway and the potential importance of primary production as a major magnification step for Hg on coral reefs is evident when concentrations of Hg in algae are compared with reef sediments. On the basis of dry- weight, turf algae had consistently higher THg and MeHg than marine sediments. At
Loa, turf algae THg was higher than in sediment by a factor of 2. At Aasu, turf algae
THg was higher than in sediment by a factor of 3-5, and for Vatia by a factor of 10-20.
For MeHg, differences between turf algae and sediment were even greater. At Loa, turf algae had 10 times higher MeHg than sediments. At Aasu, turf algae MeHg was 10-40 times higher than in sediments, and for Vatia up to 20 times higher. For Vatia and Loa, factors are estimates, based on one-half DL for calculation purposes. An enriched
MeHg:THg ratio for turf algae compared with sediments suggests the possibility of preferential sequestration of Hg in turf algae, or turf algae as a biotic methylation site in
143 the marine environment. These findings suggest that a significant proportion of Hg that accumulates in reef biota has its origin through the turf algae pathway. The uptake of
Hg directly through the primary production pathway via turf algae could potentially be of greater significance in overall Hg accumulation in upper-trophic coral reef biota than the indirect or collateral uptake of Hg from reef sediments.
4.4.2 Mercury in Reef Fish
4.4.2.1 Acanthurus lineatus
Sample logs for Acanthurus lineatus, Mullidae spp., and S. qenie, are provided in
Appendix B. For A. lineatus, all specimens collected were used for study purposes
(Aasu n=20, Vatia n=20, Loa n=20). For the Mullids, only specimens of Parupeneus cyclostomus were used for comparisons among study sites (Aasu n=5, Vatia n=5, Loa n=11) because an evaluation of inter-species variability of weight and muscle tissue Hg showed that meaningful comparisons could likely not be made with multiple species data.
There was no difference in THg and MeHg from A. lineatus among the un-impacted sites of Vatia and Aasu (p>0.05), with ~60-65 % of Hg in muscle tissue occurring as
MeHg (Table 4-24, Figure 4-18). Values for THg, MeHg, %MeHg, and weight, appear consistent for fish among the un-impacted sites. Because of the high site fidelity of these fish (Craig, 1996) and the distance between the study sites (~8 km), the probability of overlapping populations of A. lineatus between Vatia and Aasu was considered negligible. With regard to Hg in herbivorous fish, these sites were interpreted as independent, and highly representative of un-impacted coral reef systems in remote Oceania.
144 Muscle tissue THg and MeHg for A. lineatus from Loa were significantly higher
(p<0.001) than for fish from Aasu and Vatia. The six-fold increase for THg in A. lineatus from the Loa site compared with Aasu and Vatia fish, the corresponding order of magnitude increase for MeHg and the increased MeHg:THg ratio (~95%), combined with the overall smaller size of Loa fish, is not easy to explain. There was no difference in turf algae THg among the un-impacted and impacted sites, and only slight differences in turf algae MeHg among sites, so differences of Hg in food source for A. lineatus cannot be associated with differences in muscle tissue Hg among un-impacted and impacted sites, at least given the data in hand.
Table 4-24 Hg in Acanthurus lineatus muscle tissue (ng g-1 wet-weight) Weight Total Methyl Ratio Lipids Solids Sample ID (g) Hg Hg MeHg:THg (%) (%)
Loa ALOGO 1-1-1 203.1 3.402 3.662 108% 2.92 26.9 Loa ALOGO 1-1-2 165.4 4.218 3.468 82% 5.18 22.5 Loa ALOGO 1-1-3 119.8 4.320 3.903 90% 1.47 21.2 Loa ALOGO 1-1-4 130.8 6.025 4.992 83% 2.92 23.2 Loa ALOGO 1-1-5 146.8 4.246 4.259 100% 3.51 24.1
Loa ALOGO 1-2-1 135.1 5.977 4.985 83% 2.29 22.1 Loa ALOGO 1-2-2 163.0 4.444 4.202 95% 3.67 24.0 Loa ALOGO 1-2-3 130.3 3.683 3.965 108% 1.88 20.9 Loa ALOGO 1-2-4 150.3 4.008 3.632 91% 3.24 23.7 Loa ALOGO 1-2-5 138.7 13.49 14.50 107% 1.42 24.2
Loa ALOGO 2-1-1 134.1 6.285 5.148 82% 1.39 22.3 Loa ALOGO 2-1-2 165.1 5.871 5.058 86% 3.51 24.1 Loa ALOGO 2-1-3 186.5 7.596 6.738 89% 3.56 25.0 Loa ALOGO 2-1-4 175.2 6.170 6.406 104% 1.04 21.4 Loa ALOGO 2-1-5 181.2 9.772 9.936 102% 1.84 22.5
Loa ALOGO 2-2-1 183.5 4.528 4.913 109% 2.89 24.1 Loa ALOGO 2-2-2 186.4 7.352 6.695 91% 3.51 24.6 Loa ALOGO 2-2-3 202.7 5.946 5.629 95% 6.89 28.9 Loa ALOGO 2-2-4 187.7 5.086 4.539 89% 2.90 23.8 Loa ALOGO 2-2-5 174.5 5.911 5.295 90% 2.53 23.6
Mean: 163.0 ±5.7 5.92 ±0.53 5.60 ±0.57 94.1% 2.93 23.7 Range: (119.8-203.1) (3.40-13.5) (3.47-14.5) ------
Vatia ALOGO 1-1-1 329.5 0.849 0.473 56% 0.88 21.9 Vatia ALOGO 1-1-2 223.5 1.028 0.502 49% 1.91 22.4 Vatia ALOGO 1-1-3 273.2 0.845 0.309 37% 1.87 22.9 Vatia ALOGO 1-1-4 244.9 0.752 0.436 58% 0.83 21.2 Vatia ALOGO 1-1-5 228.6 0.841 0.466 55% 1.27 22.0
Vatia ALOGO 1-2-1 223.8 0.830 0.552 66% 0.92 21.2 Vatia ALOGO 1-2-2 239.3 1.300 0.635 49% 1.37 21.4
145 Table 4-24 Hg in Acanthurus lineatus muscle tissue (ng g-1 wet-weight) Weight Total Methyl Ratio Lipids Solids Sample ID (g) Hg Hg MeHg:THg (%) (%)
Vatia ALOGO 1-2-3 233.7 1.163 0.635 55% 1.55 21.7 Vatia ALOGO 1-2-4 220.1 1.410 0.440 31% 0.68 20.7 Vatia ALOGO 1-2-5 278.1 1.479 1.069 72% 0.83 21.1
Vatia ALOGO 2-1-1 230.2 0.735 0.737 100% 0.91 21.8 Vatia ALOGO 2-1-2 216.6 0.840 0.766 91% 1.55 22.4 Vatia ALOGO 2-1-3 226.7 1.183 0.928 78% 0.78 20.4 Vatia ALOGO 2-1-4 216.0 1.189 0.951 80% 1.46 21.5 Vatia ALOGO 2-1-5 249.7 0.828 0.535 65% 0.91 20.8
Vatia ALOGO 2-2-1 249.8 1.217 0.901 74% 0.86 21.2 Vatia ALOGO 2-2-2 252.5 1.012 0.775 77% 0.83 20.8 Vatia ALOGO 2-2-3 295.8 1.273 0.690 54% 1.13 22.7 Vatia ALOGO 2-2-4 277.9 1.229 1.006 82% 1.24 22.3 Vatia ALOGO 2-2-5 246.3 1.075 0.830 77% 0.69 20.6
Mean: 247.8 ±6.6 1.05 ±0.05 0.682 ±0.049 65.3% 1.12 21.6 Range: (216.0-329.5) (0.735-1.48) (0.309-1.07) ------
Aasu ALOGO 1-1-1 248.4 0.925 0.967 104% 2.69 23.2 Aasu ALOGO 1-1-2 231.1 0.605 0.415 69% 2.91 23.9 Aasu ALOGO 1-1-3 207.7 0.741 0.450 61% 2.58 23.1 Aasu ALOGO 1-1-4 259.9 1.112 0.584 53% 1.72 22.9 Aasu ALOGO 1-1-5 227.8 0.794 0.427 54% 1.86 23.4
Aasu ALOGO 1-2-1 286.6 1.009 0.416 41% 1.48 22.3 Aasu ALOGO 1-2-2 168.3 0.835 0.384 46% 1.64 23.1 Aasu ALOGO 1-2-3 223.3 0.833 0.447 54% 0.78 24.1 Aasu ALOGO 1-2-4 234.8 0.932 0.376 40% 0.92 22.3 Aasu ALOGO 1-2-5 165.5 1.059 0.351 33% 0.71 21.6
Aasu ALOGO 2-1-1 224.1 0.909 0.683 75% 0.86 21.8 Aasu ALOGO 2-1-2 190.8 1.271 1.073 84% 1.73 21.8 Aasu ALOGO 2-1-3 214.1 1.607 0.947 59% 1.17 21.4 Aasu ALOGO 2-1-4 236.0 1.272 0.772 61% 0.88 20.6 Aasu ALOGO 2-1-5 243.9 1.247 0.853 68% 1.66 22.3
Aasu ALOGO 2-2-1 227.4 1.220 0.857 70% 1.44 21.7 Aasu ALOGO 2-2-2 216.1 0.986 0.672 68% 1.27 21.3 Aasu ALOGO 2-2-3 207.3 1.056 0.746 71% 1.50 21.9 Aasu ALOGO 2-2-4 226.4 1.282 0.866 68% 1.09 22.7 Aasu ALOGO 2-2-5 229.5 1.087 0.622 57% 0.91 22.4
Mean: 223.5 ±6.3 1.04 ±0.05 0.645 ±0.051 61.8% 1.49 22.4 Range: (165.5-286.6) (0.605-1.61) (0.351-1.07) ------
Sediment Hg at Loa was significantly higher than the Aasu and Vatia sites, and may offer a plausible explanation for the differences in muscle tissue Hg for A. lineatus among the Loa and un-impacted sites. Field observations confirmed that the turf algae matt entrains fine sediment (Figure 4-19), and dissection of fish showed that full stomachs from A. lineatus contained a large (un-quantified) amount of sediment material, consistent for all fish from all sites used in this study. If it is assumed that
146 scavenging of Hg from the water column deposits Hg to turf matts in a manner generally similar to deposition of Hg in basin sediments, then increased uptake of Hg
Figure 4-18 Hg in muscle tissue - Acanthurus lineatus (mean w/SE)
10.00
9.00
8.00
7.00
6.00 wet-weight) -1 5.00
4.00
3.00 Mean Hg (ng Hg Mean g 2.00 Total Hg Methyl Hg 1.00
0.00 Aasu Vatia Loa Study Site
Figure 4-19 Fine sediment accumulation at base of turf algae (observed)
147 via ingestion of entrained turf sediment while feeding could explain the higher levels of
Hg in muscle tissue of A. lineatus at Loa compared with fish from the Asau and Vatia sites.
Of interest was that A. lineatus from all sites did not show a relationship between body weight and Hg accumulation in muscle tissue (Figures 4-20 and 4-21), although accumulation was much greater in fish from Loa than from Aasu and Vatia. The data from the Aasu and Vatia sites suggest that Hg accumulation reaches a limit in fish from un-impacted environments. The Loa data suggest that if physiological functions that limit accumulation of Hg in coral reef fish, these are easily overwhelmed by an increase in Hg above normal background levels.
Figure 4-20 Total Hg in muscle tissue vs weight - Acanthurus lineatus
10.00
9.00
8.00
7.00 r 2 = 0.0001 (n = 20) 6.00
wet-weight) Loa -1 5.00 Aasu & Vatia (pooled data) 4.00
3.00
Total(ng Hg g r 2 = 0.0033 2.00 (n = 40)
1.00
0.00 100 125 150 175 200 225 250 275 300 325 350 Whole Body Weight (g)
Accumulation patterns for Hg displayed for A. lineatus were found to be similar for the benthic-feeding carnivorous Mullids from Aasu and Vatia. Mullid patterns were more remarkable in that there was a clear limit to bioaccumulation of Hg across
148 multiple Mullid species, and among body weights that varied by a factor of ~25 (see
Section 4.4.2.2).
Figure 4-21 Methyl Hg in muscle tissue vs weight - Acanthurus lineatus
10.00
9.00
8.00
7.00 r 2 = 0.0003 (n = 20)
6.00 wet-weight)
-1 Loa 5.00 Aasu & Vatia (pooled data) 4.00
3.00 2
Methyl(ng Hg g r = 0.0036 2.00 (n = 40) 1.00
0.00 100 125 150 175 200 225 250 275 300 325 350 Whole Body Weight (g)
4.4.2.2 Mullidae spp.
Mullids feed extensively on invertebrate benthic infauna (mainly worms, mollusks, and crustaceans) primarily in sediments below the lower reef margin (Wahbeh, 1992;
Myers, 1999; Kulbicki et al., 2005), and were considered to be potentially good indicators of Hg sequestration and uptake from the sediment compartment of remote coral reef environments.
Parametric (ANOVA) and non-parametric (Kruskal-Wallis) methods agreed for comparisons among sites for Hg in Mullid muscle tissue. There was no difference in
THg or MeHg in P. cyclostomus muscle tissue among un-impacted sites (p>0.05), but fish from the impacted Harbour site had significantly higher THg and MeHg (p<0.001) than fish from Aasu or Vatia, approximately an order of magnitude greater (Table 4-25,
149 Figure 4-22). Proportion of MeHg to THg was more variable for P. cyclostomus from
Loa (81-100%) than for Vatia (88-100%) and Aasu (93-100%), but did not appear to be associated with fish body weight at any study site.
Table 4-25 Hg in Parupeneus cyclostomus muscle tissue (ng g-1 wet-weight) Weight Total Methyl Ratio Lipids Solids Sample ID (g) Hg Hg Methyl:THg (%) (%)
Loa IASINA 1-1-2 274.3 185.4 168.1 91% 0.62 21.2 Loa IASINA 1-1-4 488.6 259.4 224.1 86% 1.08 22.6 Loa IASINA 1-1-5 758.0 320.8 310.1 97% 2.83 24.3 Loa IASINA 1-2-5 437.2 293.2 219.3 75% 0.43 21.7 Loa IASINA 2-1-1 430.0 179.4 166.8 93% 2.34 24.6 Loa IASINA 2-1-2 253.9 228.1 184.0 81% 0.37 22.3 Loa IASINA 2-1-4 740.1 372.5 374.5 101% 0.47 20.5 Loa IASINA 2-1-5 239.6 266.1 270.5 102% 0.61 21.7 Loa IASINA 2-2-2 706.6 364.6 383.4 105% 0.98 22.6 Loa IASINA 2-2-3 623.2 161.0 148.6 92% 1.74 23.6 Loa IASINA 2-2-4 185.2 134.3 138.2 103% 1.12 23.7
Mean: 467.0 ±64.5 251.3 ±24.6 235.2 ±26.5 93.2% 1.14 22.6 Range: (185.2-758.0) (134.3-372.5) (138.2-383.4) ------
Vatia IASINA 1-1-2 18.9 13.14 11.66 89% 0.56 21.4 Vatia IASINA 2-1-5 332.4 56.52 60.47 107% 0.52 20.8 Vatia IASINA 2-2-1 493.1 28.44 26.42 93% 1.30 22.1 Vatia IASINA 2-2-4 37.8 46.27 47.04 102% 0.35 20.8 Vatia IASINA 2-2-5 333.2 19.15 16.79 88% 1.28 21.7
Mean: 243.1 ±92.5 32.7 ±8.17 32.5 ±9.25 95.6% 0.80 21.3 Range: (18.9-493.1) (13.1-56.5) (11.7-60.5) ------
Aasu IASINA 1-1-1 499.9 40.62 44.73 110% 2.99 25.0 Aasu IASINA 1-1-3 62.3 14.14 15.13 107% 0.57 21.8 Aasu IASINA 1-2-5 32.0 22.35 20.81 93% 0.64 21.7 Aasu IASINA 2-2-2 79.5 24.42 23.79 97% 0.57 20.9 Aasu IASINA 2-2-5 310.3 32.92 29.79 90% 1.97 23.2
Mean: 196.8 ±90.5 26.9 ±4.55 26.8 ±5.06 99.6% 1.35 22.5 Range: (32.0-499.9) (14.1-40.6) (15.1-44.7) ------
The increase in Hg in muscle tissue for P. cyclostomus by a factor of ~7-9 for the
Loa site compared with fish from Aasu and Vatia, was not greatly different than the factor of ~6 found for A. lineatus among sites.
150
Figure 4-22 Hg in Muscle tissue - Parupeneus cyclostomus (mean w/SE)
300
250
200 wet-weight) -1 150
Total Hg 100
Methyl Hg Mean Hg (ng Hg Mean g
50
0 Aasu Vatia Loa Study Site
Figure 4-23 Total Hg in muscle tissue vs weight - Parupeneus cyclostomus
400
350 r 2 = 0.44 (n = 11) 300
250 wet-weight) -1 200
150
Loa Total (ng Hg g 100 r 2 = 0.15 Aasu & Vatia (pooled data) (n = 10) 50
0 0 100 200 300 400 500 600 700 800 Whole Body Weight (g)
151 Figure 4-24 Methyl Hg in muscle tissue vs weight - Parupeneus cyclostomus
400
350 r 2 = 0.46 (n = 11) 300
250
wet-weight) -1 200
150
Loa 100 2 Aasu & Vatia (pooled data) Methyl(ng Hg g r = 0.14 (n = 10) 50
0 0 100 200 300 400 500 600 700 800 Whole Body Weight (g)
Results from Loa compared with un-impacted sites indicate that sediments are an important factor in Hg uptake in coral reef; directly for benthic feeding carnivores, and collaterally for turf algae grazers exposed to entrained fine sediments.
Four other species of Mullids collected for this study (P. multifasciatus, n=19; P. crassilabris, n=8; P. barberinus, n=4; Mulloidichthys vanicolensis, n=8) showed patterns of Hg accumulation similar to P. cyclostomus, though the number of specimens from each species and among sites were too few for rigorous statistical interpretations.
Similar to accumulation patterns seen in A. lineatus, the Mullids displayed limits to accumulation of Hg in the un-impacted habitats (Aasu and Vatia), but mechanisms that limit accumulation of Hg appeared to be overwhelmed when environmental concentrations of Hg increase marginally (Figures 4-23 and 4-24). When all Mullids collected for this study are included in the weight versus muscle tissue analysis, the pattern remains consistent (Figures 4-25 and 4-26).
152
Figure 4-25 Total Hg in muscle tissue vs weight - Mullidae spp.
400.0
350.0 r 2 = 0.25 (n = 20) 300.0
250.0 wet-weight) -1 200.0
150.0 Loa
Total(ng Hg g 100.0 r 2 = 0.19 Aasu & Vatia (pooled data) (n = 40) 50.0 Linear (Aasu & Vatia (pooled data))
0.0 0 100 200 300 400 500 600 700 800 Whole Body Weight (g)
Figure 4-26 Methyl Hg in muscle tissue vs weight - Mullidae spp.
400.0
350.0
300.0 r 2 = 0.22 (n = 20)
250.0
wet-weight) -1 200.0
150.0 Loa
100.0 2 Methyl(ng Hg g r = 0.22 Aasu & Vatia (pooled data) (n = 40) 50.0 Linear (Loa) Linear (Aasu & Vatia (pooled 0.0 data)) 0 100 200 300 400 500 600 700 800 Whole Body Weight (g)
153 Age was not considered in the accumulation patterns for A. lineatus or the Mullids, because otolith analyses were beyond the means of this study. Length, which is often correlated with Hg accumulation (as a surrogate for age) was also not considered to describe accumulation patterns in A. lineatus and the Mullids. Length is generally only weakly correlated with age for coral reef fish, and for many reef fish, adult size is attained in the first year or two, then remains relatively unchanged over a life span that may reach 40 years (Myers, 1999; Randall, 2005), so size may not be a good indicator of age. Age may be important in Hg accumulation in these lower-trophic reef fish, but there is no known data on Hg accumulation and age in coral reef fish from which to draw inferences.
Given the observed patterns, it is not likely that age is a major determining factor in
Hg accumulation in the reef fish studied here. There is a consistent pattern among both groups of fish that some of the largest fish had muscle tissue Hg at the lower end of the range, and in contrast, some of the smallest fish had some of the highest levels of Hg in muscle tissue. This is notable in P. cyclostomus from the un-impacted Aasu and Vatia sites, where body weight differed by a factor of more than 20, but less so in A. lineatus from these sites, where weight range was limited to a factor of ~2. It is improbable that
40 randomly selected fish from each group at these sites would show the consistent observed patterns for an upper limit of Hg accumulation, if there was a significant age- accumulation relationship in these fish.
4.4.2.3 Sphyraena qenie
Muscle tissue Hg from the top-trophic carnivore Sphyraena qenie (blackfin barracuda) showed a consistent pattern of increased concentration with body weight, unlike reef fish from lower trophic levels (Table 4-26, Figure 4-27), and was expectedly
154 higher than for A. lineatus and the Mullids. The magnitude of difference, however, was surprising, given the relatively small size of these predator fish, and the pristine environment where they were captured. The larger fish greatly exceeded the US EPA screening level of 0.3 parts-per-million (US EPA, 2000), and as discussed later, could warrant a public health concern for consumption, if popular in the local diet.
Table 4-26 Hg in muscle tissue for Sphyraena qenie (ng g-1 wet-weight) Weight Total Methyl Ratio Lipids Solids Sample ID (g) Hg Hg MeHg:THg (%) (%)
Bar 01 9050 741 752 101% 0.62 22.3
Bar 02 7470 650 643 99% 0.43 23.5
Bar 03 3570 105 105 100% 1.03 24.7
Figure 4-27 Hg in muscle tissue - Sphyraena qenie
1000
900 Total Hg r 2 = 0.98 Methyl Hg r 2 = 0.98 800
700
600
wet-weight) 500 -1 Total Hg 400 Methyl Hg
Hgg (ng 300
200
100
0 0 2000 4000 6000 8000 10000 Whole Body Weight (g)
155 Regression analysis showed a strong relationship between Hg in muscle tissue and whole body weight for S. qenie (r2=0.98), but was not shown to be statistically significant (p ~0.08). This was likely due to small sample size, since this relationship is well documented among top predator fish, and is indicated by the results.
Results for Hg in S. qenie are important primarily as an indicator of the Hg flux in the remote Tutuila aquatic environment. Public exposure to Sphyraenids in American
Samoa is minimal. The Sphyraenids are not significant in the artisanal fisheries of
Tutuila because their speed and wariness preclude capture by net or spear. Sphyraenids are taken frequently on a trolled lure by sport-fishing craft, but the numbers of coastal- going sport craft that troll in American Samoa are few, and barracuda are not a preferred sport fish and are generally released (P. Peshut, pers. obs., 2001-2008).
At the top trophic level of near-shore marine habitats, Sphyraenids are potentially good indicators of Hg prevalence in remote ecosystems, and other large top predators on coral reefs such as jacks and trevallys (Carangidae) and snappers (Lutjanidae) may show similar accumulation patterns. Conclusions from this limited data set are that Hg in these relatively small reef predators from un-impacted waters is quite high by world health standards, and that there is no attributable source of Hg except for background environmental levels of Hg in this remote southern hemisphere location.
4.4.3 Bio-Accumulation Factors
Bio-accumulation factors (BAFs) were calculated for marine water » turf algae » muscle tissue for A. lineatus, and marine water » muscle tissue for the Mullids and S. qenie (Tables 4-27 and 4-28). Factors calculated for MeHg should be interpreted with caution. The data, at the limit of analytical capabilities (generally ≤5 times DL), is considered sound, but an indeterminable degree of uncertainty is acknowledged.
156 Table 4-27 BAFs for Total Hg (referenced to marine water) Location Turf Algae A. lineatus P. cyclostomus S. qenie1
Aasu 104.6 103.5 104.9 ----
Vatia 104.5 103.5 105.0 ----
Loa 104.5 104.6 105.8 ----
Coastal ------106.4
1Bar01 (see Appendix B); calculation based on THg in Aasu and Vatia marine water (mean)
Table 4-28 BAFs for Methyl Hg (referenced to marine water) Location Turf Algae A. lineatus P. cyclostomus S. qenie1
Aasu 104.6 104.4 106.0 ----
Vatia 103.8 104.3 106.0 ----
Loa 104.0 105.0 106.7 ----
Coastal ------107.4
1Bar01 (see Appendix B); calculation based on MeHg in Aasu and Vatia marine water (mean)
BAFs for turf algae were similar for THg and MeHg, and for MeHg, about an order of magnitude less than reported values for phytoplankton (~105.5), the nearest counterpart to turf algae found in the literature for comparison (Mason et al., 1995). As noted earlier, if the DL for MeHg in turf algae for Vatia is used for BAF calculation rather than ½ DL as reported and used by convention, then the BAF is 104.1 for MeHg in Vatia turf. No data was found in the literature for BAFs to compare with A. lineatus and the Mullids. BAFs for S. qenie agreed very well with the range of 106-107 reported for top-trophic fish. The increase of approximately one order of magnitude for MeHg
BAFs compared with THg BAFs for A. lneatus, P. cyclostomus, and S. qenie, agrees
157 well with studies that show greater trophic transfer efficiency for MeHg than THg (e.g.,
Mason et al., 1995).
An interesting finding was the reduction of THg in muscle tissue of A. lineatus from the un-impacted sites, by an order of magnitude compared with turf algae, and that this reduction was not evident for THg at the Loa site. There is close agreement for MeHg
BAFs for A. lineatus and turf algae for Aasu and Vatia, but an order of magnitude increase for the Loa site. There were no references found in the literature to explain these patterns.
Studies on trace metals and algae show that passive and active transport of dissolved
Hg species across cellular membranes is the prevalent mechanism for metals uptake, and is shown to be particularly efficient for divalent metals, with little or no release back to the water column once metals are incorporated in tissue (Morris and Bale, 1975;
Eide et al., 1980). Uptake of Hg from the water column by diffusion across cellular membranes in phytoplankton and diatoms has been established as the principal magnification step in aquatic food webs, ~105, (Mason et al., 1995; Mason et al., 1996), and similarly, turf algae-associated Hg is probably limited to cellular uptake from the water column (Morris and Bale, 1975; Eide et al., 1980).
4.5 Human Health and Mercury in Tutuila Coral Reef Fish
Among the general population, health risks from exposure to Hg are attributed principally to MeHg through seafood consumption (e.g., US EPA, 1997; ATSDR, 1999;
Mergler et al., 2007). Mercury species are not known to be carcinogenic (ATSDR,
1999; US EPA, 2000) so non-cancer health endpoints, primarily with regard to neurotoxicity (Dales, 1972; Chang, 1977; Castoldi et al., 2001) were used as the basis for risk assessment.
158 Consumption limits in fish-meals per month, based on MeHg concentration in fish tissue, were prepared in accordance with US EPA (2000) guidance (Table 4-29). As public health determinations, consumption limits are typically conservative, a common practice in the public health fields with regard to human health and exposure to environmental pollutants (ATSDR, 1999; US EPA, 2000). Consumption limits based on national or regional standards have recognized limitations, and the applicability of values presented in Table 4-29 in terms of local population factors or preferences could not be accounted for in this study. Among the Samoan population of Tutuila, assumptions for body weight and serving size used to prepare Table 4-29 may not accurately reflect population parameters or typical consumption practices. Uncertainty cannot be avoided when estimating levels of safe exposure for environmental contaminants (ATSDR, 1999; US EPA, 2000; Tran et al., 2004), and there is no reliable population data for American Samoa to use as a basis to modify table values (Peshut and Brooks, 2005).
Table 4-29 Consumption limits for MeHg in fish tissue Muscle tissue MeHg Fish meals per month (ng g-1 wet-weight)
un-restricted (>16) ≤ 29 16 >29 - 59 12 >59 - 78 8 >78 - 120 7 134 6 156 5 190 4 230 3 310 2 470 1 940 ½ 1900 none (< ½) >1900
Adapted from Peshut and Brooks (2005); prepared in accordance with US EPA (2000) method
159 In the case of the Tutuila population, an under-estimation of average body weight would lead to an over-estimation of risk, and the conservative aspect of consumption limits would be enhanced, that is, overly conservative. Conversely, an under-estimation of serving size per meal for the local population would tend to under-estimate risk, and the conservative aspect of table values would be tend to be relaxed, that is, less conservative. With these considerations duly acknowledged, consumption limits were prepared in accordance with peer-reviewed standard practice (US EPA, 2000), and were used as the basis for consumption rates.
An advantage of this research was that consumption rate tables (Tables 4-30 and 4-
31) were based on actual MeHg concentration in tissue, whereas the consumption limit values (Table 4-29) are based on using THg concentrations in tissue, with the assumption that most or all Hg is present as MeHg (US EPA, 2000). The practice of using THg and then applying assumptions for MeHg is normally applied in the interests of expediency and costs.
Using experimentally-derived MeHg values would tend to improve the usefulness of the standard table values, by mitigating at least one potential source of uncertainty. It is notable that if speciation analysis had not been used for Hg in muscle tissue for this study, consumption limits based on US EPA (2000) guidance would considerably over- estimate risk for A. lineatus from un-impacted reefs, where MeHg in muscle tissue occurred only as ~65% of THg. A marginal over-estimation of risk would have been calculated for the Mullids from the un-impacted sites (MeHg ~90-95% of Hg).
Assumptions for 100% MeHg based consumption limits did not affect the top-trophic S. qenie, or the Mullids from Pago Pago Harbour, where MeHg occurred as ~100% of
THg.
160 Implications from these findings are that consumption limit tables based on THg residues in tissues can be overly conservative for lower-trophic fish, especially for remote locations. Speciation analysis for chemical contaminants that have varying toxicity depending on chemical form has been shown to mitigate overly conservative human health determinations (Liu et al., 2006; Yokel et al., 2006; Peshut et al., 2008).
The assessment provided here is intended to be interpreted only as an indication of the potential for impacts from Hg on human health in the remote oceanic location of
Tutuila, based on limited data, and is not a fully developed risk assessment effort.
Calculated consumption limits (Table 4-29) and allowable consumption rates (Tables 4-
30 and 4-31) differ from risk assessments in that the former are based on cause-effect relationships and quantifiable environmental parameters, whereas the latter are conclusions derived from judgment based on knowledge of local populations, and environmental and socio-economic conditions (US EPA, 2000).
Table 4-30 Consumption rates for un-impacted artisanal fishery Consumption Location Muscle tissue MeHg (fish meals Species (specimens) (ng g-1 wet-weight) per month)
A. lineatus Aasu (n=20) 0.645 un-restricted Vatia (n=20) 0.682 un-restricted
M. vanicolensis Aasu (n=3) 20.5 un-restricted Vatia (n=3) 11.3 un-restricted
P. barberinus Aasu (n=1) 53.6 16
P. crassilabris Aasu (n=4) 33.4 16 Vatia (n=1) 60.4 12
P. cyclostomus Aasu (n=5) 26.8 un-restricted Vatia (n=5) 32.5 16
P. multifasciatus Aasu (n=7) 22.2 un-restricted Vatia (n=11) 35.4 16
S. qenie Coastal (n=1) 752 1 Coastal (n=1) 643 1 Coastal (n=1) 105 8
161 Table 4-31 Consumption rates for impacted artisanal fishery Consumption Muscle tissue MeHg (fish meals Species Location (ng g-1 wet-weight) per month)
A. lineatus Loa (n=20) 5.60 un-restricted
M. vanicolensis Loa (n=2) 169 5
P. barberinus Loa (n=3) 72.6 12
P. crassilabris Loa (n=3) 239 4
P. cyclostomus Loa (n=11) 235 4
P. multifasciatus Loa (n=1) 233 4
This assessment was based on a subjectively selected consumption rate threshold of ≤
8 fish meals per month, consistent with the previous risk assessment for American
Samoa (Peshut and Brooks, 2005).
At the un-impacted sites of Aasu and Vatia, there was no basis to advise the public to limit consumption of surgeonfish or goatfish (Tables 4-30 and 4-31), based on the limited data from this study. Surprisingly, there was also no basis for an advisory for A. lineatus in Pago Pago Harbour (Table 4-31). Based on Hg residues in muscle tissue for
S. qenie, consumption for the two largest fish was limited to 1 meal per month.
Extrapolating from results for S. qenie suggest that fish >15 kg would have a recommendation of “no consumption”. Results for the top-predator S. qenie clearly indicated that public health concerns for consumption of top-trophic reef-fish may be warranted, but as discussed below, this evidence must remain indicative at this time.
Expectedly, the roving carnivorous goatfish from Pago Pago Harbour were assigned limited consumption status, consistent with previous findings for the Harbour (Peshut and Brooks, 2005).
Overall, results for reef fish from coastal reefs of Aasu and Vatia, and the Harbour reef at Loa, are consistent with conclusions from Peshut and Brooks (2005), that open
162 coastal reefs of Tutuila do not warrant fish advisories, and that the fish advisory currently in place for Pago Pago Harbour is supported.
These tables present encouraging results for the artisanal reef fish fisheries for
Tutuila, because consumption of economically and culturally important species does not appear to be of concern. In contrast, the high levels of Hg accumulated in the relatively small barracuda from un-impacted reefs, supports concerns for the persistence and bio- accumulation of Hg in un-impacted environments of remote locations, as a result of global proliferation of Hg pollution.
4.6 Hg in the Remote Tropical Environment of Tutuila
Atmospheric wet deposition on Tutuila did not differ significantly from comparable locations in the northern hemisphere, and modeling studies were shown to consistently under-estimate the Hg flux for the Tutuila region of the remote southern hemisphere.
The inter-hemispheric gradient of gaseous Hg appears to have little effect on Hg deposition rates at widely spaced locations in the Pacific hemisphere, and it is indeed an important question as to whether such a gradient exists. The most recent studies on the inter-hemispheric distribution of gaseous Hg are a decade old.
Research on Hg among environmental compartments on Tutuila Island support expectations, based on prevailing scientific opinion, that the principal source of Hg to
Tutuila aquatic systems is atmospheric deposition via rainfall. That rainfall is the probable source of Hg to Tutuila reefs is supported by the consistency in patterns for the association of Hg and TOC in reef sediments, and stream suspended solids, across a geographic distribution of catchments. The basaltic parent material of Tutuila is not likely a significant source of Hg to Island soils or aquatic systems.
163 Soils are known to be important sites of biological methylation of many elements by fungi, yeasts, algae, and bacteria, as well as by abiotic processes (Thayer, 2002).
Methylation of Hg in the terrestrial environment, however, is little studied, with most terrestrial interests for Hg focused on quantifying deposition and re-emissions (e.g.,
Ericksen et al., 2006; Xin et al., 2007; Kuiken et al., 2008). Little evidence was found in the literature for methylation of Hg by terrestrial organisms (Landner, 1971; Jereb et al., 2003; Wang et al., 2003), and none for tropical oceanic island systems. These findings expose intriguing areas for research, such as the terrestrial aspects of the methylation of Hg in the biogeochemical cycle, especially for tropical environments.
An additional localised source of Hg to the environment in this remote location was also indicated by study results. In recent decades, research and regulatory focus has been mainly on Hg emissions, and persistence of Hg in environmental compartments, with relatively less effort directed at reducing Hg in consumer goods or manufactured products. In remote locations, uncontrolled solid waste disposal, and the discharge of inadequately treated sewage wastes, are potential non-point sources of Hg to aquatic environments, and likely constitute an additional input of Hg besides atmospheric wet deposition. The proliferation of Hg through modern material lifestyles, and improper waste management by the local population, was strongly suggested by the contrasts in
Hg accumulations in sediments and biota from the un-impacted locations of Aasu and
Vatia, and the Loa site in Pago Pago Harbour. Mercury occurs in hundreds of general use consumer goods, household fixtures, paints, electronics, pharmaceuticals, and automobile components (US EPA, 2008; US FDA, 2008). The modern economy of
American Samoa includes a wide range of such items and it is common in American
Samoa for household wastes, demolition debris, and large items such as appliances and vehicles, to be disposed of haphazardly on un-occupied village land, despite adequate
164 solid waste services being provided by the local utility. Moisture, temperature, and halogens in marine aerosols enhance the corrosivity of the near-shore tropical marine environment, which accelerates the deterioration of wastes. Paints, papers, and metals goods are especially vulnerable to deterioration in this corrosive environment, so items such as batteries, inks, electronics and vehicles are potential sources of Hg input to
Tutuila reefs (US EPA, 1997).
Land disturbance for agricultural, residential, or other developments, is another potential anthropogenic non-point source of Hg to Tutuila reefs. In practice, land clearing in American Samoa is largely un-regulated by the Territorial government, although statutes are in place that provide for ample regulatory authority, and many government programmes to enforce or coerce compliance are in place as well. Despite legislated authority, government intervention for land regulation is only marginally effective in American Samoa. Community- and family-based governance is the traditional governance paradigm in American Samoa (Meade, 1928) and remains a strong tradition today (P. Peshut, pers. obs., 2001-2008). The authority of the centralised, western-style form of government receives limited, and selective, recognition (P. Peshut, pers. obs., 2001-2008). This is particularly relevant for issues of land use, for which the Samoan people have an especially independent regard.
Excessive soil erosion as a result of indiscriminate clearing on steep slopes and frequent heavy rains is common for populated catchments of Tutuila. Mobilisation of large quantities of terrestrial material could likely enhance mobilisation of Hg bound in soils, adding to the deposition of Hg on Tutuila reefs. The relative contribution of Hg to the
Tutuila environment through proliferation of Hg by the use and disposal of consumer goods, or via atmospheric wet deposition, is indeterminable at this time.
165 There are few studies published that give accounts for Hg in sediments of coral reefs
(Guzmán and García, 2002; Denton et al., 2006; Denton and Morrison, 2009) and none known for Hg in basin sediments of pristine reefs of remote tropical oceanic islands.
For this reason, results for Hg in Tutuila reef sediments are of limited use for assessing the relevance of Hg as a global pollutant. Indeed, levels of Hg in all Tutuila sediments studied are two orders of magnitude below the Effects Range Low (ERL) criteria of the
Sediment Quality Guidelines developed by US NOAA (1999). Nevertheless, there are advantageous features of data on Hg in Tutuila reef sediments that make this a valuable contribution to global Hg research, and which will be potentially useful to researchers at a later time. These include geographic location, pristine environments, extent of investigations (comprising ~2.5 km of coastline), adherence to ultra-trace field protocols for collection, and adherence to ultra-trace performance standards for chemical analyses. In future, data on sediment Hg from one or more of the un-impacted
Tutuila reef sites could serve as an important baseline or reference for local, regional, or global studies of Hg processes in remote aquatic environments.
It is an encouraging outcome of this work that lower-trophic fish in the remote un- impacted environment of Tutuila do not yet appear to accumulate Hg to an extent that would cause concern for consumption, notwithstanding evidence for the increased atmospheric burden of Hg and increased Hg deposition, world-wide. Surgeonfish and goatfish are a preferred food fish and dominate the small-scale commercial and village- based artisanal fisheries of the Samoa Islands, with similar preferences observed throughout the Pacific Islands (P. Peshut, pers. obs., 1995-2008). The fecundity, abundance, and relative ease of capture of these fish groups, via inexpensive hand- spear, ensure their position as an important cultural and economic resource if managed appropriately.
166 Results for the top-trophic carnivore S. qenie from the un-impacted coastal reefs of
Tutuila are less encouraging than for reef fish, with the implication that the globalisation of Hg and impacts of Hg on remote ecosystems are looming environmental issues. Bio-accumulation factors for these small reef-associated fish from remote, pristine coastal waters of the southern hemisphere were as high as any found for northern hemisphere studies, and indicate that the flux of Hg to Tutuila reefs is high, even though the flux is not apparent in other environmental compartments at this time, except rainfall. Moreover, these fish were near the limits of restrictions for consumption according to established health standards. There is an obvious need to explore the use of top-trophic coastal fish in remote locations as indicators of Hg contamination at the global scale.
Conspicuous among results from this study is that compartments commonly used as indicators of Hg contamination - sediments, water column, and mid-trophic fish - provided no evidence that Hg was of consequence in remote Oceania. Only Hg in precipitation, and Hg bio-accumulation in top-trophic fish, gave indications that concerns for the globalisation of Hg are warranted.
167 CHAPTER 5. CONCLUSIONS and RECOMMENDATIONS
5.1 Conclusions
This study demonstrated that prevailing concerns for Hg as a globally distributed and persistent environmental pollutant are valid. Renewed efforts for field studies for Hg in remote regions, to motivate actions on controls for Hg emissions world-wide, appear warranted. Mercury in rainfall on Tutuila Island in remote Oceania was similar to, or was slightly greater than, Hg in rainfall at locations in the northern hemisphere that were selected based on similarities in remote global characteristics. Results from
Tutuila indicate that Hg is well-mixed in the global atmosphere, and that Hg is distributed widely at the global scale. The inter-hemispheric concentration gradient for gaseous Hg that was observed in past decades appears irrelevant with regard to wet deposition of Hg for inter-hemispheric locations of the Pacific Ocean today. Inferences from results for Hg in Tutuila rainfall are that background levels of Hg in the southern hemisphere are not significantly different than for the northern hemisphere.
As far as can be ascertained, there has been no experimental research on the atmospheric concentration or wet deposition of Hg in remote global locations published within the current decade. results from modeling of Hg proliferation in the global environment appears to be conservative, and based on these experimental results from remote Oceania, contemporary global models appear to under-estimate Hg deposition rates for the remote southern hemisphere in the region of Tutuila Island. Obviously, much more experimental work is needed in the field in remote regions of the globe to better define global Hg distribution and persistence.
Atmospheric deposition as the principal source of Hg to un-impacted Tutuila aquatic environments was not un-equivocally demonstrated by this study, but the experimental evidence strongly supports this proposition. Other than atmospheric deposition,
168 potential sources of Hg to un-impacted Tutuila terrestrial and aquatic systems are limited, and can be ruled improbable. Regional volcanism can be discounted as a significant source of Hg to Tutuila reefs, based on the observed patterns for Hg distribution in reef sediments. It is unlikely that regional volcanism from the Vailulu’u
Seamount (200 km east) or the Tonga Trench (500 km southwest), could preferentially influence the distribution of Hg among reef locations on Tutuila in the pattern observed in this study. Active volcanic sources can therefore be reasonably ruled out as significant for inputs of Hg to Tutuila aquatic systems. The basaltic parent material of the island can also be eliminated from consideration as a significant source of Hg to
Tutuila soils or reefs. It is improbable that weathering of the basaltic base of the island could maintain the apparently elevated and relatively uniformly distributed levels of Hg found throughout Tutuila soils.
The strong association of Hg with organic carbon in stream suspended solids, and marine sediments, and the progression of increased Hg and organic carbon in these compartments coincidental with catchment area, strongly support that upland soils are the principal direct source of Hg to Tutuila aquatic systems. It can be reasonably argued that the only sustained input of Hg to Tutuila’s terrestrial environment that could maintain the observed patterns of Hg distribution in soils, streams suspended material, and reef sediments, must come from a source external to the island. The only external source that could explain the observed terrestrial and aquatic conditions with respect to
Hg, is atmospheric deposition.
Elevated Hg in muscle tissue of the top-trophic reef predator Sphyraena qenie is an important indicator of the ubiquity of Hg in remote global regions, and shows that Hg bio-accumulates to levels of concern for human health even in remote un-impacted environments. It is reasonable to surmise that other large top-trophic reef predators
169 found abundantly in Tutuila reef waters, such as jacks and trevallys (Carangidae) and snappers (Lutjanidae) might show similar patterns for accumulation of Hg. The
Carangids and Lutjanids are restricted in range compared with Sphyraenids, and research on Hg accumulation in these fish groups from un-impacted sites of Tutuila could be important as baseline data to enhance our understanding of Hg bio- accumulation in remote and seemingly pristine environments.
Health risks from the consumption of surgeonfish and goatfish from the artisanal reef fish fishery in American Samoa appear negligible for coastal reefs according to risk assessment protocols. Results for un-impacted reefs from this study are consistent with conclusions for coastal reefs from previous studies in American Samoa, in that contaminant levels in reef fish at locations outside of Pago Pago Harbour do not warrant a fish advisory. An unexpected finding from this study is that the herbivorous surgeonfish A. lineatus from Pago Pago Harbour (Loa) were assigned a status of “un- restricted” consumption, similar to A. lineatus from Aasu and Vatia. Despite evidence of elevated Hg in some environmental compartments on Tutuila, this herbivorous fish species do not appear to bio-accumulate Hg to any significant extent, even on the impacted Harbour reefs.
Neither the goatfish nor surgeonfish from un-impacted reefs showed any tendency to accumulate Hg in relation to body weight, which suggests some equilibrium mechanism in lower-trophic fish metabolism that can accommodate limited background levels of
Hg. Alternatively, age may be a factor for Hg accumulation in these fishes, but since it is commonly found that there is a lack of correlation between age and weight in reef fish, this may confound attempts to explain accumulation patterns. Harbour fish, however, did indicate that the relationship between muscle tissue Hg and body weight changes with increased levels of Hg in the local environment.
170 Outside of Pago Pago Harbour, the reef fish fishery of American Samoa appears healthy in terms of Hg contamination. Unless further studies indicate risks from contaminants other than Hg, the reef fish fishery of Tutuila can be enjoyed by village residents for the important socio-economic benefits it provides, and which are an historic component of Polynesian culture.
Observations that un-impacted reefs of Tutuila are apparently healthy with regard to
Hg pollution, at least on the lower trophic levels, should not be taken as deference to the apparent international complacency on issues for control of Hg emissions, or the resistance to recognition of Hg as a global contaminant. Mercury released to the environment today constitutes a pollution legacy, shared by all. Once released to the dynamic biosphere from stable ores in natural archives, Hg does not sequester to any appreciable extent at the scale of human life expectancy. Mercury volatility and mobility in environmental media ensures that today’s Hg emissions are passed to future generations. The preponderance of evidence, accumulated since Minamata, indicates that levels of this potent neurotoxin in the transitory reservoirs of the biosphere are increasing. Elevated concentrations of Hg in rainfall on Tutuila and in muscle tissue of small top-trophic fish on pristine Tutuila reefs are consistent with the observed trends of
Hg flux world-wide. Equilibrium characteristics of the Hg flux in the environment are not well known at this time, but it is reasonable to say that continued emissions to the atmospheric reservoir will presumably be reflected in a continued rise in Hg concentrations in the biosphere reservoirs, over time. Indications that lower-trophic fish in remote pristine areas of the globe are not yet significantly impacted by anthropogenic
Hg emissions are encouraging findings, for now. Encouraging findings notwithstanding, the levels of Hg in Tutuila rainfall and barracuda should serve as a signal that the global proliferation of Hg is a looming environmental issue.
171 5.2 Recommendations
Tutuila Island presents an advantageous opportunity to establish a long-term atmospheric Hg monitoring station in a remote oceanic location of the southern hemisphere. Facilities and personnel for atmospheric monitoring are currently in place at Cape Matatula on Tutuila’s easternmost point, operated by the Global Monitoring
Division (GMD) of the US National Oceanic and Atmospheric Administration. The
“clean air” station at Cape Matatula was established because of its advantageous geographic position for monitoring background trace atmospheric constituents, and is one of five similarly placed stations between the Arctic and Antarctic for coordinated atmospheric monitoring at the global scale (Barrow, Alaska, 71o N 157o W; Trinidad
Head, California, 41o N 157o W; Mauna Loa, Hawaii, 19o N 156o W; American Samoa,
14o S 171o W; South Pole, 89.9o S 25o W).
The US NOAA supports cooperative efforts with scientific investigators for atmospheric studies. A long-term monitoring programme for atmospheric TGM and Hg wet deposition could be established relatively easily and inexpensively at the Cape
Matatula site. For TGM, a range of reliable and accurate, automated and semi- automated field equipment for continuous Hg monitoring are commercially available at moderate costs. These facilities require minimal resources for installation, calibration, and startup, and only infrequent maintenance. For monitoring Hg wet deposition, abundant rainfall at Cape Matatula (~2000 mm yr-1) would facilitate manual, short- period deployment of collection equipment throughout the year, similar to the procedures implemented for this study at the Alega site. Manually deployed and retrieved sampling apparatuses, over short deployment periods, limit the potential for confounding trace Hg contamination compared with long-period deployments or from the use of un-attended automated sampling equipment.
172 A collateral benefit for an Hg monitoring initiative at Cape Matatula is that depth of rainfall and other atmospheric parameters important for Hg research are monitored continuously as fundamental components of many of the on-going programs.
Logistically, Tutuila Island has adequate communications, shipping, and transportation infrastructure to support a long-term atmospheric Hg monitoring program. Tutuila
Island presents un-precedented advantages as a global watch site for long-term trends in atmospheric Hg in the remote southern hemisphere.
As an adjunct to long-term monitoring of atmospheric Hg on Tutuila, a top-trophic fish monitoring programme would provide valuable data on how the un-impacted aquatic biological system on Tutuila responds to atmospheric Hg over time. Long-term monitoring programmes for various environmental parameters are currently in place for many Tutuila reefs, under the auspices of several territorial government agencies, and a monitoring programme for Hg accumulation in top-trophic fish could be designed and implemented efficiently as a component of existing programmes. Since biological compartments generally respond more slowly in the expression of environmental contaminants than do abiotic compartments, field collections for the reef predator monitoring programme could be on the scale of every 3-5 years.
Long-term studies for Hg in the atmosphere and biosphere from the remote global location of Tutuila Island will prove useful as one baseline from which to assess the global proliferation of Hg, answering the need widely expressed that increased experimental data from remote global locations is necessary to raise awareness for the importance of Hg as a global contaminant.
173 CHAPTER 6. REFERENCES
Adey, W.H., Goertemiller, T. 1987. Coral reef algal turfs: master producers in nutrient poor seas. Phycologia 26, 374-386.
Adey, W.H., Steneck, R.S. 1985. Highly productive eastern Caribbean reefs: synergistic effects of biological, chemical, physical, and geological factors. In: Reaka,
M.L. (ed.). The ecology of coral reefs: results of a workshop held by the American
Society of Zoologists, Philadelphia, USA. US NOAA Symposium Series, Undersea
Research 3, 163-187.
AECOS. 1991. A preliminary toxics scan of water, sediment and fish tissues from
Inner Pago Pago Harbor in American Samoa. American Samoa Environmental
Protection Agency, Pago Pago.
Aller, R.C. 1994. Bioturbation and remineralization of sedimentary organic matter: effects of redox oscillation. Chemical Geology 114, 331-345.
Andersson, A. 1979. Mercury in soils. In: Nriagu, J.O. (ed.). The biogeochemistry of mercury in the environment. Elsevier Biomedical Press, Amsterdam.
ASTM. 1982. Method D4129-82. Carbon, total organic in soil. American Society for
Testing and Materials, New York.
174 ATSDR. 1999. Toxicological Profile for Mercury. US Department of Health and
Human Services, Agency for Toxic Substances and Disease Registry, Atlanta, Georgia.
Bashkin, V.N., Howarth, R.W. 2002. Modern biogeochemistry. Kluwer, Dordrecht.
Benoit, J.M., Fitzgerald, W.F., Damman, A.W.H. 1998. The biogeochemistry of an ombrotrophic bog: evaluation of use as an archive of atmospheric deposition.
Environmental Research 78, 118-133.
Benoit, J.M., Gilmour, C.C., Mason, R.P., Heyes, A. 1999. Sulfide controls on mercury speciation and bioavailability to methylating bacteria in sediment pore waters.
Environmental Science and Technology 33, 951-957.
Benoit, J.M., Gilmour, C.C., Mason, R.P., Riedel, G.S., Riedel, G.F. 1998. Behavior of mercury in the Patuxent River estuary. Biogeochemistry 40, 249-265.
Bergan, T., Gallardo, L., Rodhe, H. 1999. Mercury in the global troposphere: a three dimensional model study. Atmospheric Environment 33, 1575-1585.
Biester, H., Kilian, R., Franzen, C., Woda, C., Mangini, A., Schöler, H.F. 2002.
Elevated mercury accumulation in a peat bog of the Magellanic Moorlands, Chile (53o
S) - an anthropogenic signal form the Southern Hemisphere. Earth and Planetary
Science Letters 201, 609-620.
175 Bligh, E.G., Dyer, W.J. 1959. A rapid method of total lipid extraction and purification.
Canadian Journal of Biochemistry and Physiology 37, 911-917.
Bloom, N.S. 1989. Determination of picogram levels of methylmercury by aqueous phase ethylation, followed by cryogenic gas chromatography with cold vapour atomic fluorescence detection. Canadian Journal of Fish and Aquatic Sciences 46, 1131-1140.
Bloom, N.S. 1992. On the chemical form of mercury in edible fish and marine invertebrate tissue. Canadian Journal of Fish and Aquatic Sciences 49, 1010-1017.
Bloom, N.S., Colman, J.A., Barber, L. 1997. Artifact formation of methyl mercury during aqueous distillation and alternative techniques for the extraction of methyl mercury from environmental samples. Fresenius Journal of Analytical Chemistry 358,
371-377.
Bloom, N.S., Crecilius, E.A. 1983. Determination of mercury in seawater at sub- nanogram per liter levels. Marine Chemistry 14, 49-59.
Bloom, N.S., Gill, G.A., Cappellino, S., Dobbs, C., McShea, L., Driscoll, C., Mason, R.,
Rudd, J. 1999. Speciation and cycling of mercury in Lavaca Bay, Texas, sediments.
Environmental Science and Technology 33, 7-13.
Bloom, N.S., Watras, C.J. 1989. Observations of methylmercury in precipitation.
Science of the Total Environment 87/88, 199-207.
176 Bodek, I., Lyman, W.J., Reehl, W.F., Rosenblatt, D.H. (eds.). 1988. Environmental inorganic chemistry. Properties, processes, and estimation methods. Pergamon Press,
New York.
Boening, D.W. 2000. Ecological effects, transport, and fate of mercury: a general review. Chemosphere 40, 1335-1351.
Bullock, O.R. 2000. Modeling assessment of transport and deposition patterns of anthropogenic mercury air emissions in the United States and Canada. Science of the
Total Environment 259, 145-157.
Bullock, O.R., Brehme, K.A. 2002. Atmospheric mercury simulation using the CMAQ model: formulation description and analysis of wet deposition results. Atmospheric
Environment 36, 2135-2146.
Cameron, E.M., Jonasson, I.R. 1972. Mercury in Precambrian shales of he Canadian
Shield. Geochimica et Cosmochimica Acta 30, 985-1005.
Carlson, R.W., (ed.). 2005. The mantle and core. Treatise on geochemistry, Volume 2.
Elsevier, Dordrecht.
Carpenter, R.C. 1985. Relationship between primary production and irradiance on coral reef algal communities. Limnology and Oceanography 30, 784-793.
177 Carpenter, R.C. 1986. Partitioning herbivory and its effects on coral reef algal communities. Ecological Monographs 56, 345-363.
Castoldi, A.F., Coccini, T., Ceccatelli, S., Manzo, L. 2001. Neurotoxicity and molecular effects of methylmercury. Brain Research Bulletin 55, 197-203.
Celo, V., Lean, D.R.S., Scott, S.L. 2006. Abiotic methylation of mercury in the aqutic environment. Science of the Total Environment 368, 126-137.
CH2M Hill. 2007. Receiving water quality monitoring, Pago Pago Harbour, non- tradewind season. American Samoa Environmental Protection Agency, Pago Pago.
Chang, L.W. 1977. Neurotoxic effects of mercury - a review. Environmental Research
14, 329-373.
Choe, K-Y., gill, G.A., Lehman, R. 2003. Distribution of particulate, colloidal, and dissolved mercury in San Francisco Bay estuary. 1. Total mercury. Limnology and
Oceanography 48, 1535-1546.
Chouvelon, T., Warnau, M., Churlaud, C., Bustamante, P. 2009. Hg concentrations and related risk assessment in coral reef crustaceans, mollusks, and fish from New
Caledonia. Environmental Pollution 157, 331-340.
178 Clarkson, T.W., Magos, L., Myers, G. 2003. Current concepts: The toxicology of mercury - current exposures and clinical manifestations. New England Journal of
Medicine 349, 1731-1737.
Compeau, G., Bartha, R. 1984. Methylation and demethylation of mercury under controlled redox, pH, and salinity conditions. Applied and Environmental
Microbiology 48, 1203-1207.
Costa, S.L., Peshut, P.J., Goldstein, C.L., Glatzel, K.A. 2007. Sediment toxicity assessment for Pago Pago Harbor. In: Proceedings of the fourth international conference on remediation of contaminated sediments, Savannah, Georgia.
Craig, P.C. 1996. Intertidal territoriality and time-budget of the surgeonfish,
Acanthurus lineatus, in American Samoa. Environmental Biology of Fishes 46, 27-36.
Craig, P.C., Choat, J.H., Axe, L.M., Saucerman, S. 1997. Population biology and harvest of the coral reef surgeonfish Acanthurus lineatus in American Samoa. Fishery
Bulletin 95, 680-693.
Dahl, A.L. 1974. The structure and dynamics of benthic algae in the coral reef ecosystem. In: Cameron, A.M. et al. (eds.). Proceedings from the 2nd International
Symposium on Coral Reefs. Great Barrier Reef Committee, Brisbane.
Dales, L.G. 1972. The neurotoxicity of alkyl mercury compounds. American Journal of Medicine 53, 219-232.
179 Daly, R.A. 1924. The geology of American Samoa. Carnegie Institute, Washington,
Publication 340.
Denton, G.R.W., Concepcion, L.P., Wood, H.R., Morrison, R.J. 2006. Trace metals in marine organisms from four harbours in Guam. Marine Pollution Bulletin 52, 1784-
1832.
Denton, G.R.W., Morrison, R.J. 2009. The impact of a rudimentary landfill on the trace metal status of Pago Bay, Guam. Marine Pollution Bulletin 58, 150-162.
DiDonato, E.M., DiDonato, G.T., Smith, L.M., Harwell, L.C., Engle, V.D., Summers,
J.K. 2007. Using the national coastal assessment methodology to evaluate the water, sediment, and fish tissue quality of American Samoa’s near-shore coastal resources.
National Park of American Samoa, Pago Pago.
D’Itri, P.A., D’Itri, F.M. 1977. Mercury contamination: a human tragedy. John Wiley and Sons, New York.
Downs, S.G., Macleod, C.L., Lester, J.N. 1998. Mercury in precipitation and its relation to bioaccumulation in fish: a literature review. Water Air and Soil Pollution
108, 149-187.
Eckman, J.E. 1985. Flow disruption by an animal-tube mimic affects sediment bacterial colonization. Journal of Marine Research 43, 419-435.
180 Eide, I., Myklestad, S., Melsom, S. 1980. Long-term uptake and release of heavy metals by Ascophyllum nodosum (Phaeophyceae) in situ. Environmental Pollution 23,
19-28.
EnviroSearch. 1994. Human health risk assessment for the consumption of fish and shellfish contaminated with heavy metals and organochlorine compounds in American
Samoa. American Samoa Environmental Protection Agency, Pago Pago.
Ericksen, J.A., Gustin, M.S., Xin, M., Weisberg, P.J., Fernandez, G.C.J. 2006. Air-soil exchange of mercury from background soils in the United States. Science of the Total
Environment 366, 851-863.
Fadini, P.S., Jardim, W.F. 2001. Is the Negro River Basin Amazon, impacted by naturally occurring mercury? Science of the Total Environment 275, 71-82.
Farber, E. 1952. The evolution of chemistry. Ronald Press, New York.
Fiji Meteorological Service. 2007. Tropical Cyclone Summary. Tropical Cyclone
Centre, Nadi.
Fitzgerald, W.F. 1989. Atmospheric and oceanic cycling of mercury. In: Riley, J.P.,
Chester, R., Duce, R.A. (eds.). Chemical Oceanography, Volume 10. SEAREX: The sea/air exchange program, 151-186.
181 Fitzgerald, W.F. 1995. Is mercury increasing in the atmosphere? The need for an atmospheric mercury network (AMNET). Water Air and Soil Pollution 80, 245-254.
Fitzgerald, W.F., Engstrom, D.R., Lamborg, C.H., Tseng, C-M., Balcom, P.H.,
Hammerschmidt, C.R. 2005. Modern and historic atmospheric mercury fluxes in northern Alaska: global sources and Arctic depletion. Environmental Science and
Technology 39, 557-568.
Fitzgerald, W.F., Engstrom, D.R., Mason, R.P., Nater, E.A. 1998. The case for atmospheric mercury contamination in remote areas. Environmental Science and
Technology 32, 1-7.
Fitzgerlad, W.F., Gill, G.A., Kim, J.P. 1984. An equatorial Pacific Ocean source of atmospheric mercury. Science 224, 597-599.
Fitzgerald, W.F., Lamborg, C.H., Hammerschmidt, C.R. 2007. Marine biogeochemical cycling of mercury. Chemical Reviews 107, 641-662.
Fitzgerald, W.F., Mason, R.P. 1996. The global mercury cycle: oceanic and anthropogenic aspects. In: Baeyens, W., Ebinghause, R., Vasiliev, O. (Eds.), Global and regional mercury cycles: sources, fluxes, and mass balances. NATO ASI Series 2.
Environment, Vol. 21. Kluwer, Dordrecht.
182 Fitzgerald, W.F., Mason, R.P., Vandal, G.M. 1991. Atmospheric cycling and air-water exchange of mercury over mid-continental lacustrine regions. Water Air and Soil
Pollution 56, 745-767.
Flanagan, F.J., Moore, R., Aruscavage, P.J. 1982. Mercury in geologic reference samples. Geostandards Newsletter 6, 25-46.
Fleming, E.J., Mack, E.E., Green, P.G., Nelson, D.G. 2006. Mercury methylation from unexpected sources: molybdate-inhibited freshwater sediments and an iron-reducing bacterium. Applied and Environmental Microbiology 72, 457-464.
Folch, J., Lees, M., Stanley, G.H.S. 1957. A simple method for the isolation and purification of total lipids from animal tissues. Journal of Biological Chemistry 226,
497-509.
Gangaiya, P., Tabudravu, J., South, R., Sotheeswaran, S. 2001. Heavy metals contamination of the Lami coastal environment, Fiji. South Pacific Journal of Natural
Sciences 19, 24-29.
GDC. 2007. Small communities wastewater facilities plan for the villages of Leloaloa,
Aua, and Onesosopo, American Samoa. American Samoa Environmental Protection
Agency, Pago Pago.
Geological Survey of Finland. 2007. FOREGS Atlas.
183 Giambelluca, T.W., Nullet, D., Nullet, M.A. 1988. Agricultural drought on south- central Pacific Islands. Professional Geographer 40, 404-415.
Gianguzza, A., Pelizzetti, E., Sammartano, S., (eds.). 2002. Chemistry of Marine
Water and Sediments. Springer-Verlag, Berlin.
Gilbert, F., Stora, G., Bertrand, J-C. 1996. In situ bioturbation and hydrocarbon fate in an experimental contaminated Mediterranean coastal ecosystem. Chemosphere 33,
1449-1458.
Gill, G.A., Bloom, N.S., Cappellino, S., Driscoll, C.T., Dobbs, C., McShea, L., Mason,
R., Rudd, J.W.M. 1999. Sediment-water fluxes of mercury in Lavaca Bay, Texas.
Environmental Science and Technology 33, 663-669.
Gill, G.A., Fitzgerald, W.F. 1985. Mercury sampling of open ocean waters at the picomolar level. Deep-Sea Research 32, 287-297.
Gill, G.A., Fitzgerald, W.F. 1987. Mercury in the surface waters of the open ocean.
Global Biogeochemical Cycles 1, 199-212.
Gill, G.A., Fitzgerald, W.F. 1987. Picomolar mercury measurements in seawater and other materials using stannous chloride reduction and two-stage gold amalgamation with gas phase detection. Marine Chemistry 20, 227-243.
184 Gill, G.A., Fitzgerald, W.F. 1988. Vertical mercury distributions in the oceans.
Geochemica et Cosmochimica Acta 52, 1719-1728.
Gilmour, C.C., Riedel, G.S., Ederington, M.C., Bell, J.T., Benoit, J.M., Gill, G.A.,
Stordal, M.C. 1998. Methylmercury concentrations and productive rates across a trophic gradient in the northern Everglades. Biogeochemistry 40, 327-345.
Givelet, N., Roos-Barraclough, F., Shotyk, W. 2003. Predominant anthropogenic sources and rates of atmospheric mercury accumulation in southern Ontario recorded by peat cores from three bogs: comparison with natural “background” values (past 8000 years). Journal of Environmental Monitoring 5, 935-949.
Gladkikh, V.P., Fedorchuk, V.P., Lisichkina, Z.N. 1975. Mercury in volcanic rocks of the association of alkaline olivinic basalts. Geokhimiya 7, 1077-1083.
Grigal, D.F. 2002. Inputs and outputs of mercury from terrestrial watersheds: a review.
Environmental Reviews 10, 1-39.
Guentzel, J.L., Landing, W.M., Gill, G.A., Pollman, C.D. 1995. Atmospheric deposition of mercury in Florida: the FAMS project (1992-1994). Water Air and Soil
Pollution 80, 393-402.
Guentzel, J.L., Landing, W.M., Gill, G.A., Pollman, C.D. 2001. Processes influencing rainfall deposition of mercury in Florida. Environmental Science and Technology 35,
863-873.
185 Gustin, M.S. 2003. Are mercury emissions from geologic sources significant? Science of the Total Environment 304, 153-167.
Gustin, M.S., Lindberg, S.E., Weisberg, P.J. 2008. An update on the natural sources and sinks of atmospheric mercury. Applied Geochemistry 23, 482-493.
Guzmán, H.M., García, E.M. 2002. Mercury levels in coral reefs along the Caribbean coast of Central America. Marine Pollution Bulletin 44, 1415-1420.
Hackney, J.M., Sze, P. 1988. Photorespiration and productivity rates of a coral reef algal turf assemblage. Marine Biology 98, 483-492.
Hammerschmidt, C.R., Fitzgerald, W.F. 2008. Sediment-water exchange of methylmercury determined from shipboard benthic flux chambers. Marine Chemistry
109, 86-97.
Harris, R., Krabbenhoft, D.P., Mason, R., Murray, M.W., Reash, R., Saltman, T. (eds.).
2007. Ecosystem response to mercury contamination. Indicators of change. CRC
Press, Boca Raton.
Harriss, R.C. 1971. Ecological implications of mercury pollution in aquatic systems.
Biological Conservation 3, 279-283.
Hay, M.E. 1991. The functional morphology of turf-forming seaweeds: persistence in stressful marine habitats. Ecology 62, 739-750.
186 Heyes, A., Mason, R.P., Kim, E-H., Sunderland, E. 2006. Mercury methylation in estuaries: insights from using measuring rates using stable mercury isotopes. Marine
Chemistry 102, 134-147.
Hollweg, T.A., Gilmour, C.C., Mason, R.P. 2009. Methylmercury production in sediments of Chesapeake Bay and the mid-Atlantic continental margin. Marine
Chemistry 114, 86-101.
Hoyer, M., Burke, J., Keeler, G. 1995. Atmospheric sources, transport and deposition of mercury in Michigan: two years of event precipitation. Water Air and Soil Pollution
80, 199-208.
Hylander, L.D., Meili, M. 2003. 500 years of mercury production: global annual inventory by region until 2000 and associated emissions. Science of the Total
Environment 304, 13-27.
Irukayama, K., Kai, F., Fujiki, M., Kondo, T. 1962. Studies on the origin of the causitive agent of Minamata disease. III. Industrial wastes containing mercury compounds from Minamata factory.
Iverfeldt, A. 1991. Occurrence and turnover of atmospheric mercury over the Nordic countries. Water Air and Soil Pollution 56, 251-265.
187 Iverson, S.J., Lang, S.L.C., Cooper, M.H. 2001. Comparison of the Bligh and Dyer and Folch methods for total lipids determinations in a broad range of marine tissues.
Lipids 36, 1283-1287.
Izuka, S.K. 1999. Summary of groundwater data for Tutuila and Aunuu, American
Samoa, for October 1987 through September 1997. US Geological Survey Open-file
Report 99-252.
Izuka, S.K., Giambelluca, T.W., Nullet, M.A. 2005. Potential evapotranspiration on
Tutuila, American Samoa. US Geological Survey Scientific Investigations Report
2005-5200.
Jackson, T.A. 1997. Long-range atmospheric transport of mercury to ecosystems, and the importance of anthropogenic emissions-a critical review and evaluation of the published evidence. Environmental Reviews 5, 99-120.
Jardim, W.F., Silvério da Silva, G. 2003. Estimation of mercury wet deposition in the tributary sub-basins of the Negro river (Amazon-Brazil) using RS/GIS tools. Journal de
Physique IV France 107, 667-670.
Jedrychowski, W., Perera, F., Jankowski, J., Rauh, V., Flak, E., Caldwell, K.L., Jones,
R.L., Pac, A., Lisowska-Miszczyk, I. 2007. Fish consumption in pregnancy, cord blood mercury level and cognitive and psychomotor development in infants followed over the first three years of life Krakow epidemiologic study. Environment
International 33, 1057-1062.
188 Jensen, S., Jerenelöv, A. 1969. Biological methylation of mercury in aquatic organisms. Nature 223, 753-754.
Jereb, V., Horvat, M., Drobne, D., Pihlar, B. 2003. Transformations of mercury in the terrestrial isopod Porcellio scaber (Crustacea). Science of the Total Environment 304,
269-284.
Jernelöv, A, Lann, H. 1971. Mercury accumulation in food chains. OIKOS 22, 403-
406.
Johansen, P., Mulvad, G., Pedersen, H.S., Hansen, J.C., Rigét, F. 2007. Human accumulation of mercury in Greenland. Science of the Total Environment 377, 173-
178.
Kainz, M., Lucotte, M., Parrish, C.C. 2003. Relationships between organic matter composition and methyl mercury content of offshore and carbon-rich littoral sediments in an oligotrophic lake. Canadian Journal of Fisheries and Aquatic Sciences 60, 888-
896.
Kannan, K., Falandysz, J. 1998. Speciation and concentration of mercury in certain coastal marine sediments. Water Air and Soil Pollution 103, 129-136.
Keeler, G.J., Landis, M.S., Norris, G.A., Christianson, E.M., Dvonch, J.T. 2006.
Sources of mercury wet deposition in eastern Ohio, USA. Environmental Science and
Technology 40, 5874-5881.
189 King, J.K., Kostka, J.E., Frisher, M.E., Saunders, F.M. 2000. Sulfate-reducing bacteria methylate mercury at variable rates in pure culture and marine sediments.
Applied Environmental Biology 66, 2430-2437.
King, J.K., Kostka, J.E., Frischer, M.E., Saunders, F.M, Jahnke, R.A. 2001. A quantitative relationship that demonstrates mercury methylation rates in marine sediments are based on the community composition and activity of sulfate-reducing bacteria. Environmental Science and Technology 35, 2491-2496.
King, J.K., Saunders, F.M., Lee, R.F., Jahnke, R.A. 1999. Coupling mercury methylation rate to sulfate reduction rates in marine sediments. Environmental
Toxicology and Chemistry 18, 1362-1369.
Kitamura, M., Sumino, K., Taira, M. 1971. Synthesis and decomposition of organomercury compounds by micro-organisms. Japan Journal of Hygiene 26, 36.
Klumpp, D.W., McKinnon, A.D. 1989. Temporal and spatial patterns in primary production of a coral-reef epilithic algal community. Journal of Experimental Marine
Biology and Ecology 131, 1-22.
Klumpp, D.W., McKinnon, A.D.; Daniel, P. 1987. Damselfish territories: zones of high productivity on coral reefs. Marine Ecology Progress Series 40, 41-51.
Kuiken, T., Gustin, M., Zhang, H., Lindberg, S., Sedinger, B. 2008. Mercury emission from terrestrial background surfaces in the eastern USA. Part II: Air/surface exchange
190 of mercury within forests from South Carolina to New England. Applied Geochemistry
23, 356-368.
Kulbicki, M., Bozec, Y-M., Labrosse, P., Letourneur, Y., Mou-tam, G., Wantiez, L.
2005. Diet composition of carnivorous fishes from coral reef lagoons of New
Caledonia. Aquatic Living Resources 18, 231-250.
Lai, S-O., Holsen, T.M., Hopke, P.K., Liu, P. 2007. Wet deposition of mercury at a
New York state rural site: concentrations, fluxes, and source areas. Atmospheric
Environment 41, 4337-4348.
Lamborg, C.H., Fitzgerald, W.F., Damman, A.W.H., Benoit, J.M., Balcom, P.H.,
Engstrom, D.R. 2002. Modern and historic atmospheric mercury fluxes in both hemispheres: global and regional mercury cycling implications. Global
Biogeochemical Cycles 16, 1104-1114.
Lamborg, C.H., Fitzgerald, W.F., O’Donnell, J., Torgerson, T. 2002. A non-steady state compartmental model of global-scale mercury biogeochemistry with interhemispheric atmospheric gradients. Geochimica et Cosmochemica Acta 66, 1105-
1118.
Lamborg, C.H., Rolfhus, K.R., Fitzgerald, W.F., Kim, G. 1999. The atmospheric cycling and air-sea exchange of mercury species in the south and equatorial Atlantic
Ocean. Deep-Sea Research II 46, 957-977.
191 Landing, W.M., Guentzel, J.L., Gill, G.A., Pollman, C.D. 1998. Methods for measuring mercury in rainfall and aerosols in Florida. Atmospheric Environment 32,
909-918.
Landing, W.M., Perry, Jr., J.J., Guentzel, J.L., Gill, G.A., Pollman, C.D. 1995.
Relationships between the atmospheric deposition of trace elements, major ions, and mercury in Florida: the FAMS Project (1992-1993). Water Air and Soil Pollution 80,
343-352.
Landner, L. 1971. Biochemical model for the biological methylation of mercury suggested from methylation studies in vivo with Neurospora crassa. Nature 230, 452-
454.
Larssen, T., de Wit, H.A., Wiker, M., Halse, K. 2008. Mercury budget of a small forested boreal catchment in southeast Norway. Science of the Total Environment 404,
290-296.
Laurier, F.J.G., Mason, R.P., Gill, G.A., Whalin, L. 2004. Mercury distributions in the
North Pacific Ocean-20 years of observations. Marine Chemistry 90, 3-19.
Leermakers, M., Baeyens, W., Ebinghaus, R., Kuballa, J., Kock, H.H. 1997.
Determination of atmospheric mercury during the North Sea experiment. Water Air and Soil Pollution 97, 257-263.
192 Lin, C-J., Pehkonen, S.O. 1999. The chemistry of atmospheric mercury: a review.
Atmospheric Environment 33, 2067-2079.
Lindberg, S.E., Stratton, W.J. 1998. Atmospheric mercury speciation: concentrations and behavior of reactive gaseous mercury in ambient air. Environmental Science and
Technology 32, 49-57.
Lindqvist, O. (ed). 1991. Mercury in the Swedish environment. Recent research on causes, consequences and corrective methods. Water Air and Soil Pollution 55, 1-251.
Lindqvist, O., Rodhe, H. 1985. Atmospheric mercury - a review. Tellus 37B, 136-159.
Lobban, C.S., Harrison, P.J. 1997. Seaweed ecology and physiology. Cambridge
University Press, Cambridge.
MacDonald, G.A. 1944. Petrography of the Samoan Islands. Bulletin of the
Geological Society of America 55, 1333-1362.
Marsh, D.O., Clarkson, T.W., Cox, C., Myers, G., Amin-Zaki, L., Al-Tikriti, S. 1987.
Fetal methylmercury poisoning. Relationship between concentration in single strands of maternal hair and child effects. Archives of Neurology 44, 1017-1022.
Marsh, D.O., Turner, M.D., Smith, J.C., Allen, P., Richdale, N. 1995. Fetal methylmercury study in a Peruvian fish-eating population. Neurotoxicology 16, 717-
726.
193 Mason, R.P., Fitzgerald, W.F. 1991. Mercury speciation in open ocean waters. Water
Air and Soil Pollution 56, 779-789.
Mason, R.P., Fitzgerald, W.F., Vandal, G.M. 1992. The sources and composition of mercury in Pacific Ocean rain. Journal of Atmospheric Chemistry 14, 489-500.
Mason, R.P., Lawson, N.M., Sheu, G.R. 2000. Annual and seasonal trends in mercury deposition in Maryland. Atmospheric Environment 34, 1691-1701.
Mason, R.P., Morel, F.M.M., Hemond, H.F. 1995. The role of microorganisms in elemental mercury formation in natural waters. Water Air and Soil Pollution 80, 775-
787.
Mason, R.P., Reinfelder, J.R., Morel, F.M.M. 1995. Bioaccumulation of mercury and methylmercury. Water Air and Soil Pollution 80, 915-921.
Mason, R.P., Reinfelder, J.R., Morel, F.M.M. 1996. Uptake, toxicity, and trophic transfer of mercury in a coastal diatom. Environmental Science and Technology 30,
1835-1845.
Mason., R.P., Sheu, G-R. 2002. Role of the ocean in the global mercury cycle. Global
Biogeochemical Cycles 16, 1093-1106.
194 Mauclair, C., Layshock, J., Carpi, A. 2008. Quantifying the effect of humic matter on the emission of mercury from artificial soil surfaces. Applied Geochemistry 23, 594-
601.
Maxson, P. 2005. Mercury flows in Europe and the world: the impact of decommissioned chlor-alkali plants. In: Pirrone, N, Mahaffey, K.R. (eds). Dynamics of mercury pollution on regional and global scales. Atmospheric processes and human exposure around the world. Springer, New York.
McAlpine, D., Araki, S. 1958. Minamata disease; an unusual neurological disorder caused by contaminated fish. Lancet 2, 629.
Meade, M. 1928. Coming of age in Samoa. Harper Collins, New York.
Meadows, P.S., Meadows, A. 1991. The geotechnical and geochemical implications of bioturbation in marine sedimentary ecosystems. Symposium of the Zoological Society of London 63, 157-181.
Mefford, T. 2002. Meteorological measurements. In: King, D.B., Schnell, R.C.,
Rosson, R.M., Sweet, C. (eds.). Climate Monitoring and Diagnostics Laboratory
Summary Report No. 26, 2000-2001.
Mefford, T. 2004. Meteorological measurements. In: Schnell, R.C., Buggle, A.,
Rosson, R.M. (eds.). Climate Monitoring and Diagnostics Laboratory Summary Report
No. 27, 2002-2003.
195 Mergler, D., Anderson, H.A., Hing Man Chan, L., Mahaffey, K.R., Murray, M.,
Sakamoto, M., Stern, A.H. 2007. Methylmercury exposure and health effects in humans: a worldwide concern. Ambio 36, 3-11.
Morel, F.M.M., Kraepiel, A.M.L., Amyot, M. 1998. The chemical cycle and bioaccumulation of mercury. Annual Reviews of Ecological Systems 29, 543-566.
Morris, A.W., Bale, A.J. 1975. The accumulation of cadmium, copper, manganese and zinc by Fucus vesiculosus in the Bristol Channel. Estuarine and Coastal Marine
Science 3, 153-163.
Morrison, R.J., Narauan, S.P., Gangaiya, P. 2001. Trace metals studies on Laucala
Bay, Suva, Fiji. Marine Pollution Bulletin 42, 397-404.
Morrissey, J. 1980. Community structure and zonation of macroalgae and hermatypic corals on a fringing reef flat of Magnetic Island (Queensland, Australia). Aquatic
Botany 8, 91-139.
Munthe, J., Hultberg, H., Iverfeldt, A. 1995. Mechanisms of deposition of methylmercury and mercury to coniferous forests. Water Air and Soil Pollution 80,
363-371.
Murata, K., Weihe, P., Araki, S., Budtz-Jorgensen, E., Grandjean, P. 1999. Evoked potentials in Faroese children prenatally exposed to methylmercury. Neurotoxicology and Teratology 21, 471-472.
196 Myers, G.J., Davidson, P.W., Cox, C., Shamlaye, C., Cernichiari, E., Clarkson, T.W.
2000. Twenty-seven years studying the human neurotoxicity of methylmercury exposure. Environmental Research 83, 275-285.
Myers, R.F. 1999. Micronesian reef fishes: a comprehensive guide to the coral reef fishes of Micronesia, 3rd revised and expanded edition. Coral Graphics, Barrigada,
Guam.
Nguyen, H.T., Kim, K-H., Kim, M-Y., Hong, S., Youn, Y-H., Shon, Z-H., Lee, J.S.
2007. Monitoring of atmospheric mercury at a global atmospheric watch (GAW) site on An-Myun Island, Korea. Water Air and Soil Pollution 185, 149-164.
Nriagu, J.O. (ed.). 1979. Topics in environmental health Vol. 3: The biogeochemistry of mercury in the environment. Elsevier/North-Holland Biomedical Press, Amsterdam.
Nriagu, J.O. 1989. A global assessment of natural sources of atmospheric trace metals.
Nature 338, 47-49.
Nriagu, J.O., Becker, C. 2003. Volcanic emissions of mercury to the atmosphere: global and regional inventories. Science of the Total Environment 304, 3-12.
Nriagu, J.O., Pacyna, J.M. 1988. Quantitative assessment of worldwide contamination of air, water and soils by trace metals. Nature 333, 134-139.
197 O’Neill, C.A. 2004. Mercury, risk, and justice. Environmental Law Reporter 34,
11070-11115.
Pacyna, E.G., Pacyna, J.M. 2002. Global emission of mercury from anthropogenic sources in 1995. Water Air and Soil Pollution 137, 149-165.
Pacyna, E.G., Pacyna, J.M., Steenhuisen, F., Wilson, S. 2006. Global anthropogenic mercury emission inventory for 2000. Atmospheric Environment 40, 4048-4063.
Peshut, P.J., Brooks, B.A. 2005. Tier 2 fish toxicity study. Chemical contaminants in fish and shellfish and recommended consumption limits for Territory of American
Samoa. American Samoa Environmental Protection Agency, Pago Pago.
Peshut, P.J., Morrison, R.J., Brooks, B.A. 2008. Arsenic speciation in marine fish and shellfish from American Samoa. Chemosphere 71, 484-492.
Petersen., G., Munthe, J., Bloxam, R. 1996. Numerical modeling of regional transport, chemical transformations and deposition of airborne mercury species. In: Baeyens, W.,
Ebinghaus, R., Vasiliev, O.F. (eds.). Regional and global mercury cycles, fluxes and mass balances. NATO-ASI Series, Series 2: Environment, Vol. 21, 191-217. Kluwer
Academic Publishers, Dordrecht.
Petersen, G., Munthe, J., Pleijel, K., Bloxam, R., Kumar, A.V. 1998. A comprehensive
Eulerian modeling framework for airborne mercury species: development and testing of the tropospheric chemistry module (TCM). Atmospheric Environment 32, 829-843.
198 Phillips, D.J.H. 1977. The use of biological indicator organisms to monitor trace metal pollution in marine and estuarine environments-a review. Environmental Pollution 13,
281-317.
Pirrone, N., Mahaffey, K.R. (eds.). 2006. Dynamics of mercury pollution on regional and global scales. Atmospheric processes and human exposure around the world.
Springer, New York.
Poissant, L., Dommergue, A., Ferrari, C.P. 2002. Mercury as a global pollutant.
Journal de Physique IV 12, 143-160.
Poulain, A.J., Garcia, E., Amyot, M., Campbell, P.G.C., Ariya, P.A. 2007. Mercury distribution, partitioning and speciation in coastal vs. inland High Arctic snow.
Geochimica et Cosmochimica Acta 71, 3419-3431.
PRISM Group. 2006. Oregon State University and Pacific Island Network, Inventory and Monitoring Program, United States National Park Service.
Pyle, D.M., Mather, T.A. 2003. The importance of volcanic emissions for the global atmospheric mercury cycle. Atmospheric Environment 37, 5115-5124.
Randall, J.E. 2005. Reef and shore fishes of the South Pacific: New Caledonia to
Tahiti and the Pitcairn Islands. University of Hawaii Press, Honolulu.
199 Rasmussen, P.E. 1994. Current methods of estimating atmospheric mercury fluxes in remote areas. Environmental Science and Technology 28, 2233-2241.
Renner, R. 2005. Asia pumps out more mercury than previously thought.
Environmental Science and Technology 39, 99A.
Rice, E.W., Eaton, A.D., Baird, R.B. (eds.). 2006. Standard methods for the examination of water and wastewater. American Public Health Association, American
Water Works Association, Water Environment Federation.
Rigét, F., Dietz, R., Born, E.W., Sonne, C., Hobson, K.A. 2007. Temporal trends of mercury in marine biota of west and northwest Greenland. Marine Pollution Bulletin
54, 72-80.
Sakamoto, M., Nakano, A., Kinjo, Y., Higashi, H., Futatsuka, M. 1991. Present mercury levels in red blood cells of nearby inhabitants about 30 years after the outbreak of Minamata disease. Ecotoxicology and Environmental Safety 22, 58-66.
Sakata, M., Asakura, K. 2007. Estimating contribution of precipitation scavenging of atmospheric particulate mercury wet deposition in Japan. Atmospheric Environment
41, 1669-1680.
Sakata, M., Marumoto, K. 2005. Wet and Dry deposition fluxes of mercury in Japan.
Atmospheric Environment 39, 3139-3146.
200 Sakata, M., Marumoto, K., Narukawa, M., Asakura, K. 2006. Regional variations in wet and dry deposition fluxes of trace elements in Japan. Atmospheric Environment 40,
521-531.
Sanders, R.D., Coale, K.H., Gill, G.A., Andrews, A.H., Stephenson, M. 2008. Recent increase in atmospheric deposition of mercury to California aquatic systems inferred from a 300-year geochronological assessment of lake sediments. Applied
Geochemistry 23, 399-407.
Santschi, P.H. 1982. Applications of enclosures to the study of ocean chemistry. In:
Grice, G.D., Reeve, M.R. (eds.). Marine mesocosms: biological and chemical research in experimental ecosystems. Springer.
Schroeder, W.H., Munthe, J. 1998. Atmospheric mercury-an overview. Atmospheric Environment 32, 809-822.
Schroeder, W.H., Munthe, J., Lindqvist, O. 1989. Cycling of mercury between water, air, and soil compartments of the environment. Water Air and Soil Pollution 48, 337.
Schuster, P.F., Krabbenhoft, D.P., Naftz, D.L., Cecil, L.D., Olson, M.L., Dewild, J.F.,
Susong, D.D., Green, J.R., Abbot, M.L. 2002. Atmospheric mercury deposition during the last 270 years: a glacial ice core record of natural and anthropogenic sources.
Environmental Science and Technology 36, 2303-2310.
Selin, M.E., Jacob, D.J., Yantosca, R.M., Strode, S., Jaeglé, L., Sunderland, E.M. 2008.
Global 3-D land-ocean-atmosphere model for mercury: present-day versus preindustrial
201 cycles and anthropogenic enrichment factors for deposition. Global Biogeochemical
Cycles 22, 2011-2024.
Selin, N.E., Selin, H. 2006. Global politics of mercury pollution: the need for multi- scale governance. Review of European Community and International Environmental
Law 15, 258-269.
Semkin, R.G., Mierle, G., Neureuther, R.J. 2005. Hydrochemistry and mercury cycling in a High Arctic watershed. Science of the Total Environment 342, 199-221.
Semu, E., Singh, B.R., Selmer-Olsen, A.R. 1986. Adsorption of mercury compounds by tropical soils II. Effect of soil: solution ratio, ionic strength, pH, and organic matter.
Water Air and Soil Pollution 32, 1-10.
Shia, R-L., Seigneur, C., Pai, P., Ko, M., Sze, N.D. 1999. Global simulation of atmospheric mercury concentrations and deposition fluxes. Journal of Geophysical
Research 104, 23747-23760.
Slemr, F., Brunke, E-G., Ebinghaus, R., Temme, C., Munthe, J., Wangberg, I.,
Schroeder, W., Steffen, A. 2003. Worldwide trend of atmospheric mercury since 1977.
Geophysical Research Letters 30, 1516-1519.
Slemr, F., Junkermann, W., Schmidt, R.W.H., Sladkovic, R. 1995. Indication of change in global and regional trends in atmospheric mercury concentrations.
Geophysical Research Letters 22, 2143-2146.
202 Slemr, F., Langer, E. 1992. Increase in atmospheric global concentrations of mercury inferred from measurements over the Atlantic Ocean. Nature 355, 434-437.
Slemr, F., Schuster, G., Seiler, W. 1985. Distribution, speciation, and budget of atmospheric mercury. Journal of Atmospheric Chemistry 3, 407-434.
Smith, C.N., Kesler, S.E., Blum, J.D., Rytuba, J.J. 2008. Isotope geochemistry of mercury in source rocks, mineral deposits and spring deposits of the California Coast
Ranges, USA. Earth and Planetary Science Letters 269, 399-407.
Sokal, R.R., Rohlf, F.J. 1981. Biometry. The principles and practices of statistics in biological research (2nd ed.). W.H. Freeman, New York.
St. Louis, V.L., Rudd, J.W.M., Kelly, C.A., Barrie, L.A. 1995. Wet deposition of methyl mercury in northwestern Ontario compared to other geographic locations.
Water Air and Soil Pollution 80, 405-414.
Staudigel, H., Hart, S.R., Pile, A., Bailey, B.E., Baker, E.T., Brooke, S., Connelly, D.P.,
Haucke, L., German, C.R., Hudson, I., Jones, D., Koppers, A.A.P., Konter, J., Lee, R.,
Pietsh, T.W., Tebo, B.M., Templeton, A.S., Zierenberg, R., Young, C.M. 2006.
Vailulu'u seamount, Samoa: Life and death on an active submarine volcano.
Proceedings of the National Academy of Sciences of the United States of America 103,
6448-6453.
203 Stearns, H.T. 1944. Geology of the Samoan Islands. Bulletin of the Geological
Society of America 55, 1279-1332.
Steding, D.J., Flegal, A.R. 2002. Mercury concentrations in coastal California precipitation: evidence of local and trans-pacific fluxes of mercury to North America.
Journal of Geophysical Research 107, 4764-4770.
Stein, E.D., Cohen, Y., Winer, A.M. 1996. Environmental distribution and transformation of mercury compounds. Critical Reviews in Environmental Science and
Technology 26, 1-43.
Steinnes, E., Sjøbakk, T.E. 2005. Order-of-magnitude increase of Hg in Norwegian peat profiles since the outset of industrial activity in Europe. Environmental Pollution
137, 365-370.
Stordal, M.C., Gill, G.A., Wen, L-S., Santschi, P.H. 1996. Mercury phase speciation in the surface waters of three Texas estuaries: importance of colloidal forms. Limnology and Oceanography 41, 52-61.
Sunderland, E.M., Cohen, M.D., Selin, N.E., Chmura, G.L. 2008. Reconciling models and measurements to assess trends in atmospheric mercury deposition. Environmental
Pollution 156, 526-535.
204 Sunderland, E.M., Gobas, F.A.P.C., Branfiruen, B.A., Heyes, A. 2006. Environmental controls on the speciation and distribution of mercury in coastal sediments. Marine
Chemistry 102, 111-123.
Swain, E.B., Jakus, P.M., Rice, G., Lupi, F., Maxson, P.A., Pacyna, J.M., Penn, A.,
Spiegel, S.J., Velga, M.M. 2007. Socioeconomic consequences of mercury use and pollution. Ambio36, 45-61.
Takaoka, S., Kawakami, Y., Fujino, T., Oh-ishi, F., Motokura, F., Kumagai, Y.,
Miyaoka, T. 2008. Somatosensory disturbance by mehtylmercury esposure.
Environmental Research 107, 6-19.
Temme, C., Slemr, F., Ebinghaus, R., Einax, J.W. 2003. Distribution of mercury over the Atlantic Ocean in 1996 and 1999-2001. Atmospheric Environment 37, 1889-1897.
Thayer, J.S. 2002. Biological methylation of less-studied elements. Applied
Organometallic Chemistry 16, 677-691.
Tran, N.L., Barraj, L., Smith, K., Javier, A., Burke, T.A. 2004. Combining food frequency and survey data to quantify long-term dietary exposure: a methyl mercury case study. Risk Analysis 24, 19-30.
Tsubaki, T., Irukayama, K. (eds.). 1977. Minamata disease. Methylmercury poisoning in Minamata and Niigata, Japan. Elsevier, Amsterdam.
205 Turekian, K.K., Wedepohl, K.H. 1961. Distribution of the elements in some major units of the Earth’s crust. Bulletin of the Geological Society of America 72, 175-192.
UNEP. 2002. Global mercury assessment. United Nations Environment Programme, Chemicals, Geneva.
UNIDO. 2007a. Report on the policy and governance initiative: Enhancing multi- stakeholder approaches to address mercury, small-scale gold mining and the instutional dynamics of change. United Nations Industrial Development Organisation. Project
EG/GLO/01/G34.
UNIDO. 2007b. Report to the UNEP Governing Council meeting: Global impacts of mercury supply and demand in small-scale gold mining. United Nations Industrial
Development Organisation. Project EG/GLO/01/G34.
Ure, A.M., Berrow, M.L. 1982. The elemental constituents of soils. In: Environmental Chemistry Volume 2, Specialist Periodical Report. Royal Society of Chemistry, London, 94-204.
US ACOE 1979. American Samoa water resources study. Baseline water quality survey. Army Corps of Engineers District, Honolulu.
USDA. 1983. Soil survey of American Samoa. Department of Agriculture, Soil
Conservation Service, Washington, D.C.
USDA. 2007. National Atmospheric Deposition Program/Mercury Deposition
Network. Department of Agriculture, Cooperative State Research, Education, and
206 Extension Service. NADP Program Office, Illinois State Water Survey, Champaign,
Illinois.
US Department of Commerce. 2001. United States census 2000. Economics and
Statistics Administration, Census Bureau, Washington, D.C.
US EPA. 1992. Pala Lagoon fish toxicity study, American Samoa. Environmental
Protection Agency, Region IX, San Francisco.
US EPA. 1994a. Method 200.7, revision 4.4. Determination of metals and trace elements in waters and wastes by inductively coupled plasma - atomic emissions spectrometry. Environmental Protection Agency, Office of Research and Development,
Cincinnati, Ohio.
US EPA. 1994b. Method 200.8, revision 5.4. Determination of trace elements in waters and wastes by inductively coupled plasma - mass spectrometry. Environmental
Protection Agency, Office of Research and Development, Cincinnati, Ohio.
US EPA. 1994c. Method 7471A. Mercury in solid or semisolid waste (manual cold- vapor technique). Environmental Protection Agency, Office of Research and
Development, Cincinnati, Ohio.
US EPA. 1996a. Method 1638. Determination of trace elements in ambient waters by inductively coupled plasma-mass spectrometry. Environmental Protection Agency,
Washington, D.C.
207 US EPA. 1996b. Method 1669. Sampling ambient water for trace metals at EPA
Water Quality Criteria levels. EPA-821/R-96/011. Environmental Protection Agency,
Office of Water, Washington, D.C.
US EPA. 1996c. Method 6010B, revision 2. Inductively coupled plasma-atomic emission spectrometry. Environmental Protection Agency, Office of Water,
Washington, D.C.
US EPA. 1997. Mercury study report to congress, EPA-452/R-97-003, Office of Air
Quality and Planning and Standards. Environmental Protection Agency, Office of
Research and Development, Washington, D.C.
US EPA. 2000. Guidance for Assessing Chemical Contaminant Data for Use in Fish
Advisories. Volume 2. Risk Assessment and Fish Consumption Limits. Third Edition.
EPA 823-B-00-008. Environmental Protection Agency, Office of Water, Washington
DC.
US EPA. 2001a. Method 1630. Methylmercury in water by distillation, aqueous ethylation, purge and trap, and CVAFS. EPA 821-R-01-020. Environmental Protection
Agency, Office of Water, Washington, D.C.
US EPA. 2001b. Appendix to Method 1631. Mercury in tissue, sludge, sediment, and soil by acid digestion and BrCl oxidation. EPA 821-R-01-013. Environmental
Protection Agency, Office of Water, Washington, D.C.
208 US EPA. 2002. Method 1631, revision E. Mercury in water by oxidation, purge and trap, and cold vapor atomic fluorescence spectrometry. EPA 821-R-02-019.
Environmental Protection Agency, Office of Water, Washington, D.C.
US EPA. 2007. Method 245.5, modified. Mercury in solid or semi-solid wastes. In:
Test methods for evaluating solid waste, physical/chemical methods. Environmental
Protection Agency, Office of Research and Development, Washington, D.C.
US EPA. 2008. Table of products that may contain mercury and management options.
Environmental Protection Agency, Office of Solid Waste and Emergency Response,
Washington, D.C.
US FDA. 2008. Mercury in drug and biological products. Food and Drug
Administration, Center for Drug Evaluation and Research, Washington, D.C.
US National Climatic Data Center. 2002. Cooperative summary of the day, TD 3200, period of record through 2001, Western United States and Pacific Islands.
US National Climatic Data Center. 2004. Comparative climate data. http://www1.ncdc.noaa.gov/pub/data/ccd-data.
US NOAA. 1999. Sediment quality guidelines developed for the National Status and
Trends Program. National Oceanic and Atmospheric Administration, Boulder,
Colorado.
209 US NOAA. 2005. Oceanographic processes affecting coral reefs. National Oceanic and Atmospheric Administration, Boulder, Colorado.
Van den Hoek, C., Mann, D.G., Jahns, H.M. 1995. Algae. An introduction to phycology. Cambridge University Press, Cambridge.
Van Noorden, R. 2008. Mercury exports banned. Chemistry World 5, 8.
Vandal, G.M., Fitzgerald, W.F., Boutron, C.F., Candelone, J-P. 1993. Variations in mercury deposition to Antarctica over the past 34,000 years. Nature 362, 621-623.
Vonk, J.W., Sijpesteijn, A.K. 1973. Studies on the methylation of mercuric chloride by pure cultures of bacteria and fungi. Antoine van Leeuwenhoek Journal of Microbiology and Serology 39, 505-513.
Wahbeh, M.J. 1992. The food and feeding habits of the goatfishes, Mulloides flavolineatus and Mulloides vanicolensis in the Gulf of Aqaba, Jordan.
Senckenbergiana Marit. 22, 245-254.
Wallace, G.T., Siebert, D.L., Holzkneckt, S.M., Thomas, W.H. 1982. The biogeochemical fate and toxicity of mercury in controlled experimental ecosystems.
Estuarine and Coastal Shelf Science 15, 151-182.
Wanders, J.B.W. 1977. The role of benthic algae in the shallow reef of Curacao
(Netherlands Antilles). III. The significance of grazing. Aquatic Botany 3, 357-390.
210 Wang, D., He, L., Wei, S., Feng, X. 2006. Estimation of mercury emission from different sources to atmosphere in Chongqing, China. Science of the Total
Environment 366, 722-728.
Wang, D., Shi, X., Wei, S., Feng, X. 2003. Accumulation and transformation of atmospheric mercury in soil. Science of the Total Environment 304, 209-214.
Wang, D.Y., Qing, C.L., Guo, T.Y., Guo, Y.J. 1997. Effects of humic acid on transport and transformation of mercury in soil-plant systems. Water Air and Soil Pollution 95,
35-43.
Wass, R.C., 1984. An Annotated Checklist of the Fishes of American Samoa. NOAA
Technical Report NMFS SSRF-781. National Oceanic and Atmospheric
Administration, U.S. Department of Commerce, Washington, D.C.
Wasserman, J.C., Amouroux, D., Wasserman, M.A.V., Donard, O.F.X. 2002. Mercury speciation in sediments of a tropical coastal environment. Environmental Technology
23, 899-910.
Watras, C.J., Bloom, N.S. 1992. Mercury and methylmercury in individual zooplankton: implications for bioaccumulation. Limnology and Oceanography 37,
1313-1318.
Weber, J.H. 1993. Review of possible paths for abiotic methylation of mercury(II) in the aquatic environment. Chemosphere 26, 2063-2077.
211 Weber, J.H. 1999. Volatile hydride and methyl compounds of selected elements formed in the marine environment. Marine Chemistry 65, 67-75.
Wedepohl, K.H. 1995. The composition of the continental crust. Geochimica et
Cosmochimica Acta 59, 1217-1232.
Weihe, P., Hansen, J.C., Murata, K., Debes, F., Jorgensen, P., Steuerwald, U., White,
R.F., Grandjean, P. 2002. Neurobehavioral performance of Inuit children with increased prenatal exposure to methylmercury. International Journal of Circumpolar
Health 61, 41-49.
Westöö, G. 1966. Determination of methylmercury in foodstuffs. Acta Chemica
Scandinavica 20, 2131-2137.
Williams, S.L., Carpenter, R.C. 1990. Photosynthesis/photon flux density relationships among components of coral reef algal turfs. Journal of Phycology 26, 36-40.
Wingert, E.A. 1981. A Coastal Zone Management Atlas of American Samoa.
University of Hawaii Cartographic Laboratory, Department of Geography. Heath
Printers, Seattle.
Wong, C.S.C., Duzgoren-Aydin, N.S., Aydin, A., Wong, M.H. 2006. Sources and trends of environmental mercury emissions in Asia. Science of the Total Environment
368, 649-662.
212 Wood, J.M., Kennedy, F.S., Rosen, C.G. 1968. Synthesis of methyl-mercury compounds by extracts of a methanogenic bacterium. Nature 220, 173-174.
Wright, A.C.S. 1963. Soils and land use of Western Samoa. New Zealand Department of Scientific and Industrial Research, Soil Bulletin 22.
Xin, M., Gustin, M.S. 2007. Gaseous elemental mercury exchange with low mercury containing soils: investigation of controlling factors. Applied Geochemistry 22, 1451-
1466.
Xin, M., Gustin, M., Johnson, D. 2007. Laboratory investigation of the potential for re- emission of atmospherically-derived Hg from soils. Environmental Science and
Technology 41, 4946-4951.
Yokel, R.A., Lasley, S.M., Dorman, D.C. 2006. The speciation of metals in mammals influences their toxicokinetics and toxicodynamics and therefore human health risk assessment. Journal of Toxicology and Environmental Health Part B 9, 63-85.
Zar, J.H. 1999. Biostatistical analysis (4th ed). Prentice-Hall, New Jersey.
Zhang, L., Wong, M.H. 2007. Environmental mercury contamination in China:
Sources and impacts. Environment International 33, 108-121.
Zheng, L., Liu, G., Chou, C-L. 2007. The distribution, occurrence and environmental effect of mercury in Chinese coals. Science of the Total Environment 384, 374-383.
213 APPENDIX A. LABORATORY QA/QC SUMMARIES
Table A-1A Laboratory QA/QC summary for Hg in rainfall Table A-1B Laboratory QA/QC summary for Hg in rainfall
Table A-2A Laboratory QA/QC summary for Hg in marine water Table A-2B Laboratory QA/QC summary for Hg in marine water
Table A-3A Laboratory QA/QC summary for Hg in marine sediment Table A-3B Laboratory QA/QC summary for Hg in marine sediment
Table A-4 Laboratory QA/QC summary for Elements in marine sediment
Table A-5A Laboratory QA/QC summary for THg and TOC in stream suspended solids Table A-5B Laboratory QA/QC summary for THg and TOC in stream suspended solids
Table A-6 Laboratory QA/QC summary for Hg in upland soils
Table A-7A Laboratory QA/QC summary for Hg in turf algae Table A-7B Laboratory QA/QC summary for Hg in turf algae
Table A-8A Laboratory QA/QC summary for Hg in reef fish Table A-8B Laboratory QA/QC summary for Hg in reef fish
214
APPENDIX B. REEF FISH SAMPLE LOGS
Table B-1 Sample log for Tutuila reef fish; Loa
Table B-2 Sample log for Tutuila reef fish; Vatia
Table B-3 Sample log for Tutuila reef fish; Aasu
Table B-4 Sample log for Tutuila reef fish; Southeast Coast
237
APPENDIX C. ADDITIONAL ELEMENTS (other than Hg) IN TUTUILA REEF SEDIMENTS
Table C-1 Additional Elements (other than Hg) in Tutuila reef sediments
Table C-2 Spearman Correlation Coefficients (ρ) for Elements in Tutuila reef sediments
245